Graphics and Media Lab

Projects:

Bibtex:

@inproceedings{zheniy21discovery,
    author = {Zheniy, Yang and Miao, Rebecca and Gavrishchaka, V. and Senyukova, O.},
    booktitle = {2021 4th International Conference on Information and Computer Technologies (ICICT)},
    doi = {10.1109/ICICT52872.2021.00048},
    language = {english},
    pages = {253--259},
    title = {Discovery of Hybrid Ensemble Models Resilient to Input Resolution Deterioration},
    year = {2021}
}

Abstract:

Video-quality measurement plays a critical role in the development of video-processing applications. In this paper, we show how video preprocessing can artificially increase the popular quality metric VMAF and its tuning-resistant version, VMAF NEG. We propose a pipeline that tunes processing-algorithm parameters to increase VMAF by up to 218.8\%. A subjective comparison revealed that for most preprocessing methods, a video's visual quality drops or stays unchanged. We also show that some preprocessing methods can increase VMAF NEG scores by up to 23.6\%.

Bibtex:

@article{siniukov21hacking,
    author = {Siniukov, Maksim and Antsiferova, Anastasia and Kulikov, Dmitriy and Vatolin, Dmitriy},
    journal = {ArXiv e-prints},
    language = {english},
    title = {Hacking VMAF and VMAF NEG: vulnerability to different preprocessing methods},
    year = {2021}
}

Abstract:

Video codecs have several dozen parameters that subtly affect the encoding rate, quality and size of the compressed video. Codec developers, as a rule, provide standard presets that on average yield acceptable performance for all videos, but for a given video, certain parameters may yield more efficient encoding. In this paper, we propose a new approach to predicting video codec presets to increase compression efficiency. Our effort involved collecting a new representative video-sequence dataset from Vimeo.com. An experimental evaluation showed relative bitrate decreases of 17.8\% and 7.9\%, respectively for the x264 and x265 codecs with standard options, all while maintaining quality and speed. Comparison with other methods revealed significantly faster automatic preset selection with a comparable improvement in results. Finally, our proposed content-adaptive method predicts presets that archive better performance than codec-developer presets from MSU Codec Comparison 2020 [1].

Bibtex:

@inproceedings{zvezdakov21machinelearningbased,
    author = {Zvezdakov, Sergey and Kondranin, Denis and Vatolin, Dmitriy},
    booktitle = {2021 Picture Coding Symposium (PCS)},
    doi = {10.1109/pcs50896.2021.9477507},
    language = {english},
    pages = {1--5},
    publisher = {IEEE Bristol, UK},
    title = {Machine-Learning-Based Method for Content-Adaptive Video Encoding},
    year = {2021}
}

Abstract:

Quality assessment plays a key role in creating and comparing video compression algorithms. Despite the development of a large number of new methods for assessing quality, generally accepted and well-known codecs comparisons mainly use the classical methods like PSNR, SSIM and new method VMAF. These methods can be calculated following different rules: they can use different frame-by-frame averaging techniques or different summation of color components. In this paper, a fundamental comparison of various versions of generally accepted metrics is carried out to find the most relevant and recommended versions of video quality metrics to be used in codecs comparisons. For comparison, we used a set of videos encoded with video codecs of different standards, and visual quality scores collected for the resulting set of streams since 2018 until 2021

Bibtex:

@article{antsiferova21objective,
    author = {Antsiferova, Anastasia and Yakovenko, Alexander and Safonov, Nickolay and Kulikov, Dmitriy and Gushin, Alexander and Vatolin, Dmitriy},
    journal = {ArXiv e-prints},
    language = {english},
    title = {Objective video quality metrics application to video codecs comparisons: choosing the best for subjective quality estimation},
    year = {2021}
}

Abstract:

The critical constraint of mobile devices is a limited battery life that is significantly reduced during video playback. The power efficiency of video playback mainly depends on the used compression standard, video-decoder, and device model. We propose a software-based method to estimate the power consumption of video-decoders on various Android devices. Experiments on two devices of the same model show a small variation of the power playback consumption and a lack of dependence between the power consumption and the battery level. We have implemented an automatic system that includes the VEQE Android application to measure the power consumption of decoders and a server to collect the power metrics. Our system has collected power-consumption and decoding-speed dataset for video-decoders of six standards (AV1, HEVC, VP9, H.264, VP8, and MPEG-4) operating on 285 devices, representing 147 models. We demonstrate some slices of the created dataset: the top 30 models and video-decoders in terms of power efficiency for playback and for decoding only, as well as video-decoder ratings by power consumption and decoding speed for a given device model.

Bibtex:

@inproceedings{kazantsev21power,
    author = {Kazantsev, Roman and Vatolin, Dmitriy},
    booktitle = {2021 Picture Coding Symposium (PCS)},
    doi = {10.1109/PCS50896.2021.9477481},
    language = {english},
    pages = {1--5},
    title = {Power Consumption of Video-Decoders on Various Android Devices},
    year = {2021}
}

Bibtex:

@article{gushchin21boundary,
    author = {Gushchin, Alexander and Antsiferova, Anastasia and Vatolin, Dmitriy},
    journal = {ArXiv e-prints},
    language = {english},
    title = {Shot boundary detection method based on a new extensive dataset and mixed features},
    year = {2021}
}

Abstract:

The modern realistic computer graphics is based on physically correct lighting simulation. One of the main and computationally difficult problems is the calculation of light transport or global illumination, i.e. the distribution of light in a virtual scene taking into account multiple reflections, light scattering and various interactions of light with the scene objects. This problem is studied in hundreds of books and papers. They describe dozens of computational methods and their modifications. Our survey not only lists and briefly describes them but also gives some kind of a “map” of existing works that helps the reader to find one’s bearings, understand the advantages and drawbacks of these methods and thus select an appropriate basic approach. Special attention is paid to such characteristics of the methods as their robustness and universality with respect to models, the clarity of their verification, the possibility of efficient implementation on GPUs and the constraints imposed on the scene or illumination phenomena. In contrast to existing surveys, we try to analyze not only the efficiency of the methods but their limitations and the complexity of software implementation as well. In addition, the results of the authors' own numerical experiments that illustrate some of our conclusions are presented.

Bibtex:

@article{frolov21light,
    address = {Road Town, United Kingdom},
    author = {Frolov, V. A. and Voloboy, A. G. and Ershov, S. V.},
    doi = {10.1134/S0361768821040034},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    pages = {298--326},
    publisher = {Road Town, United Kingdom},
    title = {Light Transport in Realistic Rendering: State-of-the-Art Simulation Methods},
    volume = {47},
    year = {2021}
}

Bibtex:

@inproceedings{rukhovich21iterdet,
    address = {Cham, Switzerland},
    author = {Rukhovich, Danila and Sofiiuk, Konstantin and Galeev, Danil and Barinova, Olga and Konushin, Anton},
    booktitle = {Lecture Notes in Computer Science},
    doi = {10.1007/978-3-030-73973-7\_33},
    isbn = {978-3-030-72609-6},
    language = {english},
    pages = {344--354},
    publisher = {Cham, Switzerland},
    title = {IterDet: Iterative Scheme for Object Detection in Crowded Environments},
    volume = {12602},
    year = {2021}
}

Bibtex:

@inproceedings{galeev21learning,
    address = {Cham, Switzerland},
    author = {Galeev, Danil and Sofiiuk, Konstantin and Rukhovich, Danila and Romanov, Mikhail and Barinova, Olga and Konushin, Anton},
    booktitle = {Lecture Notes in Computer Science},
    doi = {10.1007/978-3-030-73973-7\_11},
    isbn = {978-3-030-72609-6},
    language = {english},
    pages = {108--118},
    publisher = {Cham, Switzerland},
    title = {Learning High-Resolution Domain-Specific Representations with a GAN Generator},
    volume = {12602},
    year = {2021}
}

Abstract:

In 2020 the outbreak of Covid-19 influenced lives of billions of people all around the globe and motivated governments of different countries to revisit the current situation with regards to public healthcare systems and to methods used in modern medicine. As the workload on radiologists and physicians increased, so did the demand on systems that automatically analyse medical images and detect pathologies. Many current computer vision papers assume that the solution would be integrated into a healthcare system. However improvement according to “classic” metrics like mAP or IoU does not necessarily mean improvement from the radiologist’s point of view. In this paper we suggest that while calculating metrics, averaging should be performed not by all studies, but by different groups of studies, in order to be close to human perception of a quality of a segmentation. And that we should count the number of false positive components, found outside lungs, because the presence of such components is negatively perceived by radiologists. Also we propose a method that improves the segmentation of lung pathologies and pleural effusion according to the points given above.

Bibtex:

@article{lashchenova21improvement,
    address = {Road Town, United Kingdom},
    author = {Lashchenova, D. and Gromov, A. and Konushin, A. and Mesheryakova, A.},
    doi = {10.1134/S0361768821030063},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    pages = {327--333},
    publisher = {Road Town, United Kingdom},
    title = {The Improvement of Segmentation of Lung Pathologies and Pleural Effusion on CT-scans of Patients with Covid-19},
    volume = {47},
    year = {2021}
}

Abstract:

We propose a novel neural-network-based method to perform matting of videos depicting people that does not require additional user input such as trimaps. Our architecture achieves temporal stability of the resulting alpha mattes by using motion-estimation-based smoothing of image-segmentation algorithm outputs, combined with convolutional-LSTM modules on U-Net skip connections.We also propose a fake-motion algorithm that generates training clips for the video-matting network given photos with ground-truth alpha mattes and background videos. We apply random motion to photos and their mattes to simulate movement one would find in real videos and composite the result with the background clips. It lets us train a deep neural network operating on videos in an absence of a large annotated video dataset and provides ground-truth training-clip foreground optical flow for use in loss functions.

Bibtex:

@article{molodetskikh21temporally,
    author = {Molodetskikh, Ivan and Erofeev, Mikhail and Moskalenko, Andrey and Vatolin, Dmitry},
    journal = {ArXiv e-prints},
    language = {english},
    title = {Temporally Coherent Person Matting Trained on Fake-Motion Dataset},
    year = {2021}
}

Bibtex:

@inproceedings{lavrushkin21stereoscopic,
    author = {Lavrushkin, Sergey and Molodetskikh, Ivan and Kozhemyakov, Konstantin and Vatolin, Dmitriy},
    booktitle = {Electronic Imaging, 3D Measurement and Data Processing 2020},
    doi = {10.2352/issn.2470-1173.2021.2.sda-350},
    language = {english},
    title = {Stereoscopic quality assessment of 1,000 VR180 videos using 8 metrics},
    year = {2021}
}

Abstract:

Современная реалистичная компьютерная графика базируется на физически корректноммоделировании распространения света. Одной из основных и трудно вычислимых задач при этомявляется расчет глобальной освещенности, т.е. распределения света в виртуальной сцене, учитывающиймножественные отражения и рассеяния света и всевозможные виды взаимодействия его с объектамисцены. Этой проблеме посвящены сотни публикаций, описывающие десятки методов вычисленияглобальной освещенности и их модификации. В данной обзорной статье мы бы хотели не простоперечислить и кратко описать эти методы, но и дать некоторую "карту" существующих работ, котораяпозволит читателю сориентироваться, понять их достоинства и недостатки и, тем самым, выбрать длясебя подходящий базовый метод. Особое внимание уделяется таким характеристикам методов какнадёжность и универсальность в отношении используемых моделей, прозрачность их верификации,возможность эффективной реализации на GPU, а также накладываемые на сцену или феноменыосвещённости ограничения. В отличие от существующих обзорных работ анализируется не толькоэффективность методов, но также их ограничения и сложность программной реализации. Кроме того,мы предоставляем результаты собственных численных экспериментов с различными методами,служащих иллюстрациями к выводам.

Bibtex:

@article{frolov21sovremennoe,
    author = {ФРОЛОВ, В. А. and ВОЛОБОЙ, А. Г. and ЕРШОВ, С. В. and ГАЛАКТИОНОВ, В. А.},
    doi = {10.15514/ISPRAS-2021-33(2)-1},
    issn = {2220-6426; 2079-8156},
    journal = {Труды Института системного программирования РАН (электронный журнал)},
    language = {russian},
    number = {2},
    pages = {7--48},
    title = {Современное состояние методов расчёта глобальной освещённости в задачах реалистичной компьютерной графики},
    volume = {33},
    year = {2021}
}

Abstract:

Image harmonization is an important step in photo editing to achieve visual consistency in composite images by adjusting the appearances of a foreground to make it compatible with a background. Previous approaches to harmonize composites are based on training of encoder-decoder networks from scratch, which makes it challenging for a neural network to learn a high-level representation of objects. We propose a novel architecture to utilize the spaceof high-level features learned by a pre-trained classification network. We create our models as a combination of existing encoder-decoder architectures and a pre-trained foreground-aware deep high-resolution network. We extensively evaluate the proposed method on the existing imageharmonization benchmark and set up a new state-of-the-art in terms of MSE and PSNR metrics. The code and trained models are available publicly.

Bibtex:

@inproceedings{sofiiuk21foregroundaware,
    author = {Sofiiuk, Konstantin and Popenova, Polina and Konushin, Anton},
    booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)},
    language = {english},
    pages = {1620--1629},
    title = {Foreground-Aware Semantic Representations for Image Harmonization},
    year = {2021}
}

Abstract:

This paper presents a software-based method for estimating the power consumption of video decoders on various Android devices. Using this method, we developed an automatic system that consists of the VEQE Android application to measure the power consumption of video decoders and a server to collect the metrics. The system allowed us to create power-consumption and decoding-speed dataset for video decoders operating on 236 devices, representing 147 models. The paper presents several charts: the top 30 models and video decoders in terms of power efficiency for playback and for decoding only, as well as video-decoder ratings by power consumption and decoding speed for a given device model.

Bibtex:

@inproceedings{kazantsev21videodecoder,
    author = {Kazantsev, Roman and Yanushkovsky, Vladimir and Vatolin, Dmitriy},
    booktitle = {2021 Data Compression Conference (DCC)},
    language = {english},
    publisher = {IEEE Snowbird, UT, USA},
    title = {Video-Decoder Power Consumption on Android Devices: Power-Estimation Method, Dataset Creation, and Analysis Results},
    year = {2021}
}

Abstract:

Задача создания цветных версий уже существующих черно-белых изображений и видео возникла сразу же после того, как научились снимать и показывать цветные изображения и видео. Процесс добавления и распространения цвета в изображениях и видео в оттенках серого называется колоризацией. И хотя наряду с развитием методов цифровой обработки изображений и видео колоризация также постепенно автоматизируется, для ее осуществления до сих используется много ручного труда. В данной работе рассматривается задача распространения цвета в видео по одному цветному кадру, который может быть получен путем ручной колоризации либо путем применения методов автоматической колоризации изображений. Для решения этой задачи рассмотрено два типа нейросетевых методов: первые основаны на предсказании сепарабельных фильтров, вторые используют оптический поток для распространения цвета между кадрами. Большинство из рассмотренных методов показали высокое качество работы на тестовых выборках с сохранением качества при колоризации на большое количество кадров вперед.

Bibtex:

@article{lavrushkin20razrabotka,
    author = {Лаврушкин, С. В. and Ватолин, Д. С.},
    issn = {1684-2634},
    journal = {Цифровая обработка сигналов},
    language = {russian},
    number = {3},
    pages = {25--34},
    title = {Разработка нейросетевых методов распространения цвета в видео},
    year = {2020}
}

Abstract:

Shooting video in 3D format can introduce stereoscopic artifacts, potentially causing viewers visual discomfort. In this work, we consider three common stereoscopic artifacts: color mismatch, sharpness mismatch, and geometric distortion. This paper introduces two neural-network-based methods for simultaneous color- and sharpness-mismatch estimation, as well as for estimating geometric distortions. To train these networks we prepared large datasets based on frames from full-length stereoscopic movies and compared the results with methods that previously served in analyses of full-length stereoscopic movies. We used our proposed methods to analyze 100 videos in VR180 format-a new format for stereoscopic videos in virtual reality (VR). This work presents overall results for these videos along with several examples of detected problems.

Bibtex:

@inproceedings{lavrushkin20neuralnetworkbased,
    author = {Lavrushkin, Sergey and Kozhemyakov, Konstantin and Vatolin, Dmitriy},
    booktitle = {2020 International Conference on 3D Immersion (IC3D)},
    doi = {10.1109/ic3d51119.2020.9376385},
    isbn = {978-1-66540-336-8},
    language = {english},
    publisher = {IEEE},
    title = {Neural-Network-Based Detection Methods for Color, Sharpness, and Geometry Artifacts in Stereoscopic and VR180 Videos},
    year = {2020}
}

Abstract:

This paper presents a method for generating stereoscopic or multi-angle video frames using a computer game (Grand Theft Auto V). We developed a mod that captures synthetic frames allows us to create geometric distortions like those that occur in a real video. These distortions are the main cause of viewer discomfort when watching 3D movies. Datasets generated in this way can aid in solving problems related to machine-learning-based assessment of stereoscopic- or multi-angle-video quality. We trained a convolutional neural network to evaluate perspective distortions and converged camera axes in stereoscopic video, then tested it on real 3D movies. The neural network discovered multiple examples of these distortions.

Bibtex:

@inproceedings{malyshev20stereoscopic,
    author = {Malyshev, Kirill and Lavrushkin, Sergey and Vatolin, Dmitriy},
    booktitle = {2020 International Conference on 3D Immersion (IC3D)},
    doi = {10.1109/ic3d51119.2020.9376375},
    isbn = {978-1-66540-336-8},
    language = {english},
    publisher = {IEEE},
    title = {Stereoscopic Dataset from A Video Game: Detecting Converged Axes and Perspective Distortions in S3D Videos},
    year = {2020}
}

Abstract:

We consider the problem of people counting in video surveillance. This is one of the most popular tasks in video analysis, because this data can be used for predictive analytics and improvement of customer services, traffic control, etc. Our method is based on the object tracking in video with low framerate. We use the algorithm from as a baseline and propose several modifications that improve the quality of people counting. One of the main modifications is to use a head detector instead of a body detector in the tracking pipeline. Head tracking is proved to be more robust and accurate as the heads are less susceptible to occlusions. To find the intersection of a person with a signal line, we either raise the signal lines to the level of the heads or perform a regression of bodies based on the available head detections. Our experimental evaluation has demonstrated that the modified algorithm surpasses original in both accuracy and computational efficiency, showing a lower counting error on a lower detection frequency.

Bibtex:

@inproceedings{kuplyakov20distributed,
    address = {Aachen, Germany},
    author = {Kuplyakov, D. and Geraskin, Y. and Mamedov, T. and Konushin, A.},
    booktitle = {Proceedings of the 30th International Conference on Computer Graphics and Machine Vision},
    doi = {10.51130/graphicon-2020-2-3-26},
    language = {english},
    pages = {1--12},
    publisher = {Aachen, Germany},
    series = {CEUR Workshop Proceedings},
    title = {A Distributed Tracking Algorithm for Counting People in Video by Head Detection},
    volume = {2744},
    year = {2020}
}

Bibtex:

@inproceedings{chernyshev20bladder,
    address = {Aachen, Germany},
    author = {Chernyshev, Vadim and Gromov, Alexander and Konushin, Anton and Mesheryakova, Anna},
    booktitle = {Proceedings of the 30th International Conference on Computer Graphics and Machine Vision},
    doi = {10.51130/graphicon-2020-2-3-36},
    language = {english},
    pages = {1--12},
    publisher = {Aachen, Germany},
    series = {CEUR Workshop Proceedings},
    title = {Bladder Semantic Segmentation},
    volume = {2744},
    year = {2020}
}

Abstract:

Some computer vision tasks become easier with known camera calibration. We propose a method for camera focal length, location and orientation estimation by observing human poses in the scene. Weak requirements to the observed scene make the method applicable to a wide range of scenarios. Our evaluation shows that even being trained only on synthetic dataset, the proposed method outperforms known solution. Our experiments show that using only human poses as the input also allows the proposed method to calibrate dynamic visual sensors.

Bibtex:

@article{shalimova20camera,
    author = {Shalimova, E. A. and Shalnov, E. V. and Konushin, A. S.},
    doi = {10.18287/2412-6179-CO-600},
    journal = {Computer Optics},
    language = {english},
    number = {3},
    pages = {385--392},
    title = {Camera parameters estimation from pose detections},
    volume = {44},
    year = {2020}
}

Abstract:

This work is devoted to the traffic sign detection on images using deep learning methods. We focus on the problem of detector transfer to new datasets with different road signs. We present an algorithm for distilling a set of unlabelled data to select the most informative images to be labeled. This method allows to significantly reduce the amount of data labeling with a small decline of detector performance

Projects:

Bibtex:

@inproceedings{popov20distillation,
    address = {Aachen, Germany},
    author = {Popov, Alexey and Shakhuro, Vlad and Konushin, Anton},
    booktitle = {Proceedings of the 30th International Conference on Computer Graphics and Machine Vision},
    doi = {10.51130/graphicon-2020-2-3-33},
    language = {english},
    pages = {1--12},
    publisher = {Aachen, Germany},
    series = {CEUR Workshop Proceedings},
    title = {Data Distillation for Traffic Sign Detection},
    volume = {2744},
    year = {2020}
}

Abstract:

The paper considers the task of obtaining a quality assessment of facialimages for usage in various video surveillance systems, video analytics and biometric identification. Accuracy of person recognition and classification depends on the quality of the input images. We consider an approach to obtaining single face image quality assessment using neural network model, which is trained onpairs of images that are split into two possible classes: the quality of the firstimage is better or worse than the quality of the second one. Two modificationsof the selected baseline algorithm are proposed. A face recognition system is applied to change the loss function and image and face quality attributes are usedwhen training the model. Experimental studies of the proposed modificationsshow their effectiveness. The accuracy of selecting the best and worst frame isincreased by 1.3\% and 1.9\%, respectively.

Bibtex:

@inproceedings{lisin20improving,
    address = {Aachen, Germany},
    author = {Lisin, Nikita and Gromov, Alexander and Konushin, Vadim and Konushin, Anton},
    booktitle = {Proceedings of the 30th International Conference on Computer Graphics and Machine Vision},
    doi = {10.51130/graphicon-2020-2-3-28},
    language = {english},
    pages = {1--12},
    publisher = {Aachen, Germany},
    series = {CEUR Workshop Proceedings},
    title = {Improving the Neural Network Algorithm for Assessing the Quality of Facial Images},
    volume = {2744},
    year = {2020}
}

Bibtex:

@inproceedings{shvetsov20neural,
    address = {Aachen, Germany},
    author = {Shvetsov, Fedor and Konushin, Anton and Sokolova, Anna},
    booktitle = {Proceedings of the 30th International Conference on Computer Graphics and Machine Vision},
    doi = {10.51130/graphicon-2020-2-4-17},
    language = {english},
    pages = {1--12},
    publisher = {Aachen, Germany},
    series = {CEUR Workshop Proceedings},
    title = {Neural Network Model for Face Recognition from Dynamic Vision Sensor},
    volume = {2744},
    year = {2020}
}

Bibtex:

@inproceedings{nikitin20pairwise,
    address = {Aachen, Germany},
    author = {Nikitin, Mikhail and Konushin, Vadim and Konushin, Anton},
    booktitle = {Proceedings of the 30th International Conference on Computer Graphics and Machine Vision},
    doi = {10.51130/graphicon-2020-2-3-30},
    language = {english},
    pages = {1--12},
    publisher = {Aachen, Germany},
    series = {CEUR Workshop Proceedings},
    title = {Pairwise Ranking Distillation for Deep Face Recognition},
    volume = {2744},
    year = {2020}
}

Bibtex:

@inproceedings{lashchenova20segmentation,
    address = {Aachen, Germany},
    author = {Lashchenova, Daria and Gromov, Alexander and Konushin, Anton and Mesheryakova, Anna},
    booktitle = {Proceedings of the 30th International Conference on Computer Graphics and Machine Vision},
    doi = {10.51130/graphicon-2020-2-4-20},
    language = {english},
    pages = {1--12},
    publisher = {Aachen, Germany},
    series = {CEUR Workshop Proceedings},
    title = {Segmentation of Lungs, Lesions, and Lesion Types on Chest CT Scans of Patients with Covid-19},
    volume = {2744},
    year = {2020}
}

Abstract:

In recent years, deep learning-based methods for monocular visual odometry have made good progress and now demonstrate state-of-the-art results on the well-known KITTI benchmark. However, collecting ground truth camera poses for training deep visual odometry models requires special equipment and thus might be difficult and expensive. To overcome this limitation, there have been proposed a number of unsupervised methods that exploit geometric relations between depth and motion. However, there is still a large gap in accuracy between unsupervised and supervised methods. In this work, we propose a simple method for generating self-supervision for visual odometry. During training, it requires dense depth maps and an approximate motion distribution of a target platform (e.g. a car or a robot). For each input frame, we sample camera motion from the given distribution, then using a depth map we compute an optical flow that corresponds to the sampled camera motion. Then, this generated optical flow serves as an input to a visual odometry model, while the sampled camera motion serves as a ground truth output.Experiments on KITTI demonstrate that a deep visual odometry method trained in the proposed self-supervised manner outperforms unsupervised visual odometry methods, thus reducing the gap between the methods that do not require supervision and fully supervised methods. The source code is available on GitHub.

Bibtex:

@inproceedings{slinko20selfsupervised,
    author = {Slinko, Igor and Vorontsova, Anna and Zhukov, Dmitry and Barinova, Olga and Konushin, Anton},
    booktitle = {2nd International Conference on Video, Signal and Image Processing (VSIP '2020)},
    doi = {10.1145/3442705.3442706},
    language = {english},
    pages = {1--6},
    publisher = {ACM Jakarta, Indonesia. New York, NY, USA},
    title = {Self-Supervised Visual Odometry with Ego-Motion Sampling},
    year = {2020}
}

Abstract:

Deep neural networks have become a mainstream approach to interactive segmentation. As we show in our experiments, while for some images a trained network provides accurate segmentation result with just a few clicks, for some unknown objects it cannot achieve satisfactory result even with a large amount of user input. Recently proposed backpropagating refinement scheme (BRS) introduces an optimization problem for interactive segmentation that results in significantly better performance for the hard cases. At the same time, BRS requires running forward and backward pass through a deep network several times that leads to significantly increased computational budget per click compared to other methods. We propose f-BRS (feature backpropagating refinement scheme) that solves an optimization problem with respect to auxiliary variables instead of the network inputs, and requires running forward and backward passes just for a small part of a network. Experiments on GrabCut, Berkeley, DAVIS and SBD datasets set new state-of-the-art at an order of magnitude lower time per click compared to original BRS. The code and trained models are available at https://github. com/saic-vul/fbrs\_interactive\_segmentation.

Bibtex:

@inproceedings{sofiiuk20rethinking,
    author = {Sofiiuk, Konstantin and Petrov, Ilia and Barinova, Olga and Konushin, Anton},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops},
    doi = {10.1109/CVPR42600.2020.00865},
    language = {english},
    pages = {8623--8632},
    title = {f-BRS: Rethinking Backpropagating Refinement for Interactive Segmentation},
    year = {2020}
}

Abstract:

В работе рассматривается задача подсчета автомобилей в крупномасштабных системах видеонаблюдения. Применяется подход, основанный на сопровождении автомобилей и генерации событий пересечения траектории их движения с заданным сигнальным отрезком. За счет распределенности обработки и применения детекции не ко всем кадрам, а к их разряженному множеству, удается снизить требуемое число вычислительных ресурсов, увеличить производительность и получить алгоритм, способный качественно работать в реальном времени. Адаптируется и модифицируется подход, ранее предложенный для отслеживания людей [1]. Улучшение модуля оценки скорости и уточнение модели движения позволяют уменьшить частоту детекции в 3 раза. Предложенный алгоритм может работать при частоте детекции 3 Гц, сохраняя приемлемое качество.

Bibtex:

@inproceedings{shirokov20algoritm,
    author = {Широков, А. П. and Купляков, Д. А. and Конушин, А. С.},
    booktitle = {GraphiCon 2020: труды 30-й Международной конференции по компьютерной графике и машинному зрению (Санкт-Петербург 22–25 сентября 2020 г.). Том 1 / Институт прикладной математики имени М. В. Келдыша РАН. – Санкт-Петербург, 2020. – 151 с},
    language = {russian},
    pages = {100--108},
    title = {Алгоритм подсчета автомобилей в крупномасштабных системах видеонаблюдения},
    volume = {1},
    year = {2020}
}

Abstract:

В работе исследуется возможность применения нейронных сетей для классификации объектов, которых мало или совсем нет в обучающей выборке, на примере задачи классификации редких дорожных знаков. Рассматриваются нейронные сети, обученные с использованием сравнительной функции потерь и её модификациями, а также методы генерации синтетических выборок для задач классификации. В качестве базового метода используется индексирование классов объектов при помощи нейросетевых признаков. Проводится сравнение классификаторов, обученных при помощи трёх видов синтетических выборок, а также их смесей с реальными данными. Предлагается метод классификации редких дорожных знаков, использующий нейросетевой дискриминатор редких и частых знаков. Проведённая экспериментальная оценка показала, что предложенный метод позволяет классифицировать редкие дорожные знаки без существенной потери качества на частых знаках.

Projects:

Bibtex:

@article{faizov20klassifikatsija,
    author = {Фаизов, Б. В. and Шахуро, В. И. and Санжаров, В. В. and Конушин, А. С.},
    doi = {10.18287/2412-6179-co-601},
    issn = {0134-2452},
    journal = {Компьютерная оптика},
    language = {russian},
    number = {2},
    pages = {236--243},
    title = {Классификация редких дорожных знаков},
    volume = {44},
    year = {2020}
}

Bibtex:

@article{zvezdakova20bsqrate,
    address = {Road Town, United Kingdom},
    author = {Zvezdakova, Anastasia and Kulikov, Dmitriy and Zvezdakov, Sergey and Vatolin, Dmitriy},
    doi = {10.1134/S0361768820030111},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {46},
    pages = {183--194},
    publisher = {Road Town, United Kingdom},
    title = {BSQ-rate: a new approach for video-codec performance comparison and drawbacks of current solutions},
    year = {2020}
}

Bibtex:

@article{zvezdakova20bsqratenovyj,
    author = {Звездакова, Анастасия and Куликов, Дмитрий and Звездаков, Сергей and Ватолин, Дмитрий},
    doi = {10.15514/ISPRAS-2020-32(1)-5},
    journal = {Труды Института системного программирования РАН},
    language = {russian},
    number = {1},
    pages = {89--108},
    title = {BSQ-rate: новый подход к сравнению производительности видеокодеков и недостатки существующих решений},
    volume = {32},
    year = {2020}
}

Abstract:

Abstract.In recent years, the field of image inpainting has developed rapidly,learning based approaches show impressive results in the task of filling missing parts in an image. But most deep methods are strongly tied to the resolution of the images on which they were trained. A slight resolution increase leads to serious artifacts and unsatisfactory filling quality. These methods are therefore unsuitable for interactive image processing. In this article, we propose a method that solves the problem of inpainting arbitrary-size images. We also describe a way to better restore texture fragments in the filled area. For this,we propose to use information from neighboring pixels by shifting the original image in four directions. Moreover, this approach can work with existing inpainting models, making them almost resolution independent without the need for retraining. We also created a GIMP plugin that implements our technique.The plugin, code, and model weights are available at https://github.com/a-mos/HighResolutionImageInpainting

Bibtex:

@inproceedings{moskalenko20twostage,
    author = {Moskalenko, A. and Erofeev, M. and Vatolin, D.},
    booktitle = {Proceedings of the 30th International Conference on Computer Graphics and Vision},
    doi = {10.51130/graphicon-2020-2-4-18},
    language = {english},
    title = {Deep Two-Stage High-Resolution Image Inpainting},
    year = {2020}
}

Abstract:

Modern video codecs have many compression-tuning parameters from which numerous configurations (presets) can be constructed. The large number of presets complicates the search for one that delivers optimal encoding time, quality, and compressed-video size. This paper presents a machine-learning-based method that helps to solve this problem. We applied the method to the x264 video codec: it searches for optimal presets that demonstrate 9-20\% bitrate savings relative to standard x264 presets with comparable compressed-video quality and encoding time. Our method is faster upto 10 times than existing solutions.

Bibtex:

@inproceedings{kazantsev20machinelearningbased,
    author = {Kazantsev, Roman and Zvezdakov, Sergey and Vatolin, Dmitriy},
    booktitle = {2020 Data Compression Conference (DCC)},
    doi = {10.1109/dcc47342.2020.00079},
    language = {english},
    pages = {374--378},
    publisher = {IEEE Snowbird, UT, USA},
    title = {Machine-Learning-Based Method for Finding Optimal Video-Codec Configurations Using Physical Input-Video Features},
    year = {2020}
}

Abstract:

Области открытия/закрытия по движению представляют собой набор пикселей, которые видны только в одном из двух подряд идущих кадров. Поиск областей открытия/закрытия является важной задачей компьютерного зрения. Точное определение этих областей позволит улучшить качество многих методов обработки видео, таких как: интерполяция кадров, оценка оптического потока, распространение цвета и других. Большинство существующих методов основано на оптимизации функции энергии, что приводит к большим вычислительным затратам. Также стоит учитывать, что качественное нахождение областей открытия/закрытия затруднительно без какой-либо информации о движении между кадрами, но и при оценке оптического потока знание данных областей позволяет избежать ложных соответствий между пикселями кадров. Исходя из этих рассуждений в рамках данного исследования был разработан новый метод поиска областей открытия/закрытия, основанный на алгоритме построения карт оптического потока PWC-net. Ключевой идеей является построение пирамиды признаков с разными разрешениями для обработки кадров. Данный способ обработки заимствован из стандартных методов компьютерной графики и хорошо зарекомендовал себя. Также было проведено экспериментальное сравнение разработаного метода с 15 аналогами на наборе данных MPI-Sintel.

Bibtex:

@article{velikanov20nejrosetevoj,
    address = {Москва},
    author = {Великанов, М. С. and Анзина, А. Б. and Лаврушкин, С. В. and Ватолин, Д. С.},
    doi = {10.24412/FgiBpj3YL\_E},
    issn = {2307-8162},
    journal = {International Journal of Open Information Technologies},
    language = {russian},
    number = {3},
    pages = {1--7},
    publisher = {Москва},
    title = {Нейросетевой алгоритм поиска областей открытия/закрытия в видеопоследовательностях},
    volume = {8},
    year = {2020}
}

Abstract:

This paper provides a comparative study and performance analysis ofdi erent rasterization algorithms and approaches. Unlike many other papers, we don'tfocus on rasterization itself, but investigate complete graphics pipeline with 3D transfor-mations, Z-bu er, perspective correction and texturing that, on the one hand, allow usto implement a useful subset of OpenGL functionality and, on the other hand, considervarious bottlenecks in the graphics pipeline and how di erent approaches manage them.Our ultimate goal is to nd a scalable rasterizer technique that on the one hand e ectivelyuses current CPUs and on the other hand is accelerating with the extensive developmentof hardware. We explore the capabilities of scan-line and half-space algorithms raster-ization, investigate di erent memory layout for frame bu er data, study the possibilityof instruction-level and thread-level parallelism to be applied. We also study relativeeciency of di erent CPU architectures (in-order CPUs vs out-of-order CPUs) for thegraphics pipeline implementation and tested our solution with x64, ARMv7 and ARMv8instruction sets. We were able to propose an approach that could outperform highly op-timized OpenSWR rasterizer for small triangles. Finally, we conclude that despite a hugebackground high-performance software rasterization still has a lot of interesting topics forfuture research.

Bibtex:

@article{frolov20comparative,
    author = {Frolov, V. A. and Galaktionov, V. A. and Barladyan, B. H.},
    doi = {10.20948/mathmontis-2020-47-13},
    issn = {0354-2238},
    journal = {Mathematica Montisnigri},
    language = {english},
    pages = {152--175},
    title = {Comparative study of high performance software rasterization techniques},
    volume = {47},
    year = {2020}
}

Bibtex:

@article{frolov20riscv,
    author = {Фролов, В. А. and Галактионов, В. А. and Санжаров, В. В.},
    doi = {10.26295/OS.2020.53.12.002},
    journal = {Открытые системы. СУБД},
    language = {russian},
    number = {2},
    pages = {30--34},
    title = {RISC-V: стандарт, изменивший мир микропроцессоров},
    year = {2020}
}

Bibtex:

@inproceedings{sanzharov20variable,
    author = {Sanzharov, Vadim and Frolov, Vladimir and Voloboy, Alexey},
    booktitle = {CPT2020 The 8th International Scientific Conference on Computing in Physics and Technology Proceedings, Part 1},
    doi = {10.30987/conferencearticle\_5fce27723872e5.04814843},
    language = {english},
    pages = {197--201},
    publisher = {ANO "Scientific and Research Center for Information in Physics and Technique" Nizhny Novgorod},
    title = {Variable photorealistic image synthesis for training dataset generation},
    year = {2020}
}

Abstract:

Nvidia RTX – это закрытая аппаратно-ускоренная технология трассировки лучей от компании Nvidia. Поскольку детали реализации неизвестны, в сообществе разработчиков было много вопросов о том, что на самом деле представляет из себя аппаратная реализация: какие именно части в конвейере трассировки лучей ускорены аппаратно, а что может быть эффективно реализовано программно. В этой статье мы представляем результаты наших экспериментов с RTX, направленные на понимание внутренней работы этой технологии. В нашей работе мы постарались ответить на вопросы, волнующие разработчиков по всему миру: какое ускорение можно на практике получить по сравнению с программной реализацией и в чем его технологическая основа? Насколько трудоемко будет разрабатывать рендер-систему с поддержкой аппаратного ускорения, которая в то же время может работать на GPU и без RTX (т.е. реализуя трассировку лучей программно), или даже производить вычисления на CPU? Насколько эффективна программная эмуляция RTX, доступная на предыдущем поколении видеокарт Nvidia, и насколько возможно приблизить ее эффективность к аппаратной? Какова будет трудоемкость создания аналога RTX при необходимости запуска приложения на видеокартах других производителей?

Bibtex:

@article{sanzharov20issledovanie,
    author = {Санжаров, В. В. and Фролов, В. А. and Галактионов, В. А.},
    doi = {10.31857/s0132347420030061},
    issn = {0132-3474},
    journal = {Программирование},
    language = {russian},
    number = {4},
    pages = {65--72},
    title = {Исследование технологии Nvidia RTX},
    year = {2020}
}

Abstract:

Система команд - это стержень, вокруг которого строится весь остальной процессор. Ошибки или негибкость в решениях, однажды заложенные в систему команд, остаются с этим поколением процессоров навсегда. Поэтому одна из ключевых причин, по которой рост производительности современных CPU замедлился, заключается в том, что исходный код процессоров "испортился" в прямом и переносном смысле этого слова: процессоры внутри становятся сложными, из-за чего их дальнейшее развитие затрудняется. Разработка современных ЭВМ (CPU, GPU или специализированных систем) - это крайне дорогостоящий процесс, состоящий из большого количества затратных статей. Поэтому вопрос цены, или, скорее, целесообразности разработки процессора является ключевым. В данной работе мы провели исследование существующих популярных систем команд процессора и сделали выводы о перспективности в настоящее время направления RISC-V и других открытых систем команд CPU. Мы постарались ответить на следующие вопросы: почему система команд процессора - это действительно важно? Почему именно RISC-V, чем он лучше остальных? Какие возможности RISC-V открывает для российских разработчиков и какие у него есть аналоги?

Bibtex:

@article{frolov20issledovanie,
    author = {Фролов, В. А. and Галактионов, В. А. and Санжаров, В. В.},
    doi = {10.15514/ispras-2020-32(2)-7},
    issn = {2220-6426; 2079-8156},
    journal = {Труды Института системного программирования РАН (электронный журнал)},
    language = {russian},
    number = {2},
    pages = {81--98},
    title = {Исследование технологии RISC-V},
    volume = {32},
    year = {2020}
}

Abstract:

В работе предложен ряд методов, которые могут быть использованы для синтеза изображений интерьеров в целях обучения искусственного интеллекта. Предложенные методы решают проблему генерации обучающих выборок в комплексе, начиная от автоматической генерации 3D контента и заканчивая непосредственно рендерингом. Одна из основных целей разработанной системы – обеспечить достаточную производительность при генерация наборов фото-реалистичных изображений интерьеров при помощи GPU.

Bibtex:

@article{frolov20sintez,
    address = {Москва},
    author = {Фролов, В. А. and Феклисов, Е. Д. and Трофимов, М. А. and Волобой, А. Г.},
    doi = {10.20948/prepr-2020-81},
    issn = {2071-2898; 2071-2901},
    journal = {Препринты ИПМ им. М.В. Келдыша},
    language = {russian},
    number = {81},
    pages = {20--10},
    publisher = {Москва},
    title = {Синтез реалистичных изображений интерьеров для обучения нейросетей},
    year = {2020}
}

Abstract:

В данной работе предлагается подход к генерации наборов изображений путем фотореалистичного рендеринга с управляемой рандомизацией параметров 3D-сцен. Полученные таким образом наборы изображений могут быть использованы в задачах компьютерного зрения, например, для обучения моделей искусственного интеллекта. Описывается архитектура системы, реализующей предлагаемый подход, приводятся примеры применения к конкретным задачам.

Bibtex:

@article{sanzharov20sistema,
    address = {Москва},
    author = {Санжаров, В. В. and Фролов, В. А. and Волобой, А. Г. and Галактионов, В. А. and Павлов, Д. С.},
    doi = {10.20948/prepr-2020-80},
    issn = {2071-2898; 2071-2901},
    journal = {Препринты ИПМ им. М.В. Келдыша},
    language = {russian},
    number = {80},
    pages = {1--29},
    publisher = {Москва},
    title = {Система генерации наборов изображений для задач компьютерного зрения на основе фотореалистичного рендеринга},
    year = {2020}
}

Abstract:

The problem of tracking people in a video stream with the aim of counting them is studied. Modern video surveillance systems, such as the Moscow video surveillance system, use hundreds of thousands of cameras. The use of modern methods developed for working on a single computer with an expensive graphical processor is economically inefficient for such large-scale systems. In this paper, a distributed tracking algorithm is proposed. It makes it possible to reduce the amount of computational resources due to detecting people in a sparse set of frames. The detection is performed on servers installed in a data center, while the video stream is processed by local camera computation nodes. The experimental evaluation showed that the proposed algorithm provides acceptable quality at the detection rate of 4/3 Hz.

Bibtex:

@article{kuplyakov19distributed,
    address = {Road Town, United Kingdom},
    author = {Kuplyakov, D. A. and Shalnov, E. V. and Konushin, V. S. and Konushin, A. S.},
    doi = {10.1134/s0361768819040042},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {4},
    pages = {163--170},
    publisher = {Road Town, United Kingdom},
    title = {A Distributed Tracking Algorithm for Counting People in Video},
    volume = {45},
    year = {2019}
}

Abstract:

A method for the semi-automatic collection of samples for learning face identification algorithms is proposed. In the experimental evaluation, the operation of the face identification algorithm on ethnically diverse data is considered. The algorithm operation is also evaluated on the data with a wide variation of ages. The proposed method makes it possible to expand the training sample by indexing new data.

Bibtex:

@article{bagrov19semiautomatic,
    address = {Road Town, United Kingdom},
    author = {Bagrov, N. Yu and Konushin, A. S. and Konushin, V. S.},
    doi = {10.1134/S0361768819030022},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {3},
    pages = {133--139},
    publisher = {Road Town, United Kingdom},
    title = {A Semi-Automatic Method of Collecting Samples for Learning a Face Identification Algorithm},
    volume = {45},
    year = {2019}
}

Abstract:

We present Adaptive Instance Selection network architecture for class-agnostic instance segmentation. Given an input image and a point (x, y), it generates a mask for the object located at (x, y). The network adapts to the input point with a help of AdaIN layers [??], thus producing different masks for different objects on the same image. AdaptIS generates pixel-accurate object masks, therefore it accurately segments objects of complex shape or severely occluded ones. AdaptIS can be easily combined with standard semantic segmentation pipeline to perform panoptic segmentation. To illustrate the idea, we perform experiments on a challenging toy problem with difficult occlusions. Then we extensively evaluate the method on panoptic segmentation benchmarks. We obtain state-of-the-art results on Cityscapes and Mapillary even without pretraining on COCO, and show competitive results on a challenging COCO dataset. The source code of the method and the trained models are available at https://github.com/saic-vul/adaptis

Bibtex:

@inproceedings{sofiiuk19adaptis,
    author = {Sofiiuk, Konstantin and Barinova, Olga and Konushin, Anton},
    booktitle = {The IEEE International Conference on Computer Vision (ICCV)},
    doi = {10.1109/ICCV.2019.00745},
    language = {english},
    pages = {7355--7363},
    title = {AdaptIS: Adaptive Instance Selection Network},
    year = {2019}
}

Abstract:

We present a novel dataset for training and benchmarking semantic SLAM methods. The dataset consists of 200 long sequences, each one containing 3000-5000 data frames. We generate the sequences using realistic home layouts. For that we sample trajectories that simulate motions of a simple home robot, and then render the frames along the trajectories. Each data frame contains a) RGB images generated using physically-based rendering, b) simulated depth measurements, c) simulated IMU readings and d) ground truth occupancy grid of a house. Our dataset serves a wider range of purposes compared to existing datasets and is the first large-scale benchmark focused on the mapping component of SLAM. The dataset is split into train/validation/test parts sampled from different sets of virtual houses. We present benchmarking results forboth classical geometry-based and recent learning-based SLAM algorithms, a baseline mapping method, semantic segmentation and panoptic segmentation.

Bibtex:

@inproceedings{kirsanov19discoman,
    author = {Kirsanov, Pavel and Gaskarov, Airat and Konokhov, Filipp and Sofiiuk, Konstantin and Vorontsova, Anna and Slinko, Igor and Zhukov, Dmitry and Bykov, Sergey and Barinova, Olga and Konushin, Anton},
    booktitle = {2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
    doi = {10.1109/IROS40897.2019.8967921},
    language = {english},
    pages = {2470--2477},
    title = {DISCOMAN: Dataset of Indoor SCenes for Odometry, Mapping and Navigation},
    year = {2019}
}

Abstract:

Monocular depth estimation is the task of obtaining a measure of distance for each pixel using a single image. It is an important problem in computer vision and is usually solved using neural networks. Though recent works in this area have shown significant improvement in accuracy, the state-of-the-art methods tend to require massive amounts of memory and time to process an image. The main purpose of this work is to improve the performance of the latest solutions with no decrease in accuracy. To this end, we introduce the Double Refinement Network architecture. The proposed method achieves state-of-the-art results on the standard benchmark RGB-D dataset NYU Depth v2, while its frames per second rate is significantly higher (up to 18 times speedup per image at batch size 1) and the RAM usage is lower.

Bibtex:

@inproceedings{durasov19double,
    author = {Durasov, Nikita and Romanov, Mikhail and Bubnova, Valeriya and Bogomolov, Pavel and Konushin, Anton},
    booktitle = {2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
    doi = {10.1109/IROS40897.2019.8968227},
    language = {english},
    pages = {5889--5894},
    title = {Double Refinement Network for Efficient Indoor Monocular Depth Estimation},
    year = {2019}
}

Abstract:

Layout estimation is a challenge of segmenting a cluttered room image into floor, walls and ceiling. We applied Double refinement network proved to be efficient in the depth estimation to generate heat maps for room key points and edges. Our method is the first not using encoder-decoder architecture for the room layout estimation. ResNet50 was utilized as a backbone for the network instead of VGG16 commonly used for the task, allowing the network to be more compact and faster. We designed a special layout score function and layout ranking algorithm for key points and edges output. Our method achieved the lowest pixel and corner errors on the LSUN data set. The input image resolution is 224*224.

Bibtex:

@inproceedings{kruzhilov19double,
    author = {Kruzhilov, Ivan and Romanov, Mikhail and Babichev, Dmitry and Konushin, Anton},
    booktitle = {Palaiahnakote S., Sanniti di Baja G., Wang L., Yan W. (eds) Pattern Recognition. ACPR 2019},
    doi = {10.1007/978-3-030-41404-7\_39},
    language = {english},
    pages = {557--568},
    publisher = {Springer},
    series = {Lecture Notes in Computer Science},
    title = {Double Refinement Network for Room Layout Estimation},
    volume = {12046},
    year = {2019}
}

Abstract:

Nowadays face recognition systems are widely used in the world. In China these systems are used in safe cities projects in production, in Russia they are used mostly in closed-loop systems like factories, business centers with biometric access control or stadiums. Closed loop means that we need to identify people from a fixed dataset: in factory it’s a list of employees, in stadium it’s a list of ticket owners. The most challenging task is to identify people from some large city with an open dataset: we don’t have a fixed set of people in the city, it’s rapidly changing due to migration. Another limit is the accuracy of the system: we can’t make a lot of false positive errors (when a person is incorrectly recognized as another person) because number of human operators is limited and they are expensive. We propose an approach to maximize face recognition accuracy for a fixed false positive error rate using limited amount of hardware.

Bibtex:

@article{bagrov19recognition,
    address = {Netherlands},
    author = {Bagrov, N. U. and Konushin, A. S. and Konushin, V. S.},
    doi = {10.5194/isprs-archives-XLII-2-W12-11-2019},
    issn = {0924-2716},
    journal = {ISPRS Journal of Photogrammetry and Remote Sensing},
    language = {english},
    pages = {11--15},
    publisher = {Netherlands},
    title = {FACE RECOGNITION WITH LOW FALSE POSITIVE ERROR RATE},
    year = {2019}
}

Abstract:

Modern biometric systems based on face recognition demonstrate high recognition quality, but they are vulnerable to face presentation attacks, such as photo or replay attack. Existing face anti-spoofing methods are mostly based on texture analysis and due to lack of training data either use hand-crafted features or fine-tuned pretrained deep models. In this paper we present a novel CNN-based approach for face anti-spoofing, based on joint analysis of the presence of a spoofing medium and eye blinking. For training our classifiers we propose the procedure of synthetic data generation which allows us to train powerful deep models from scratch. Experimental analysis on the challenging datasets (CASIA-FASD, NUUA Imposter) shows that our method can obtain state-of-the-art results.

Bibtex:

@article{nikitin19antispoofing,
    author = {Nikitin, M. and Konushin, V. and Konushin, A.},
    doi = {10.18287/2412-6179-2019-43-4-618-626},
    journal = {Computer Optics},
    language = {english},
    number = {4},
    pages = {618--626},
    title = {Face anti-spoofing with joint spoofing medium detection and eye blinking analysis},
    volume = {43},
    year = {2019}
}

Abstract:

In this paper, we consider the task of statistical modeling of 3D human shape and pose. Current advances in computer graphics and 3D scanning and reconstruction technologies established new application areas for parametrical human body models. These urge the need both in highly detailed human body model that can qualitatively represent details of human appearance, and in a framework for fine-tuning this model using newly collected data. We contribute in both of these aspects in our work by presenting a framework for fully automatic creation of highly detailed human body model based on Skinned Multi-Person Linear (SMPL) human body model. The key features of our framework are mesh subdivision technique that increases the granularity of the model, modified non-rigid deformation algorithm (NRD) for smooth and precise registration of 3D scans and weighted registration process that allows controlling registration in low confidence areas of the 3D scan (holes, artifacts and voluminous haircuts). We propose and evaluate two body models with different detail level and show that even low detailed model outperforms existing body models in terms of registration accuracy and cumulative relative variance.

Bibtex:

@inproceedings{patakin19finetuning,
    address = {New York, N.Y., United States},
    author = {Patakin, Nikolay and Petrov, Ilia and Guzov, Vladimir and Konushin, Anton},
    booktitle = {VSIP 2019: 2019 International Conference on Video, Signal and Image Processing},
    doi = {10.1145/3369318.3369335},
    language = {english},
    pages = {21--27},
    publisher = {New York, N.Y., United States},
    title = {Fine-tuning SMPL: A Framework for Highly Detailed Statistical Human Model Building},
    year = {2019}
}

Abstract:

Gait recognition is a computer vision problem complementing other identification problems such as face or iris recognition. Unlike other identifiers, gait is based on the motion of body points and thus it can be captured by Dynamic Vision Sensor (DVS). In this work, we explore the possibility of gait recognition in the event stream by visualizing it and applying the existing method which achieves state-of-the-art results on several benchmarks. During the investigations several related problems such as moving object detection and human pose estimation are considered as auxiliary ones. Several algorithm settings are evaluated and compared. The obtained results show that all the problems can be solved in event-based data with high quality, which is close to the quality achieved on conventional colored videos.

Bibtex:

@inproceedings{sokolova19human,
    author = {Sokolova, Anna and Konushin, Anton},
    booktitle = {2019 16th International Conference on Machine Vision Applications (MVA)},
    doi = {10.23919/mva.2019.8758019},
    language = {english},
    pages = {1--6},
    publisher = {IEEE},
    series = {IEEE Xplore Digital Library},
    title = {Human identification by gait from event-based camera},
    year = {2019}
}

Abstract:

Simultaneous localisation and mapping (SLAM) is an essential component of robotic systems. In this work we perform a feasibility study of RGB-D SLAM for the task of indoor robot navigation. Recent visual SLAM methods, e.g. ORBSLAM2 [9], demonstrate really impressive accuracy, but the experiments in the papers are usually conducted on just a few sequences, that makes it difficult to reason about the robustness of the methods. Another problem is that all available RGB-D datasets contain the trajectories with very complex camera motions. In this work we extensively evaluate ORBSLAM2 to better understand the state-of-the-art. First, we conduct experiments on the popular publicly available datasets for RGB-D SLAM across the conventional metrics. We perform statistical analysis of the results and find correlations between the metrics and the attributes of the trajectories. Then, we introduce a new large and diverse HomeRobot dataset where we model the motions of a simple home robot. Our dataset is created using physically-based rendering with realistic lighting and contains the scenes composed by human designers. It includes thousands of sequences, that is two orders of magnitude greater than in previous works. We find that while in many cases the accuracy of SLAM is very good, the robustness is still an issue.

Bibtex:

@inproceedings{prokhorov19measuring,
    author = {Prokhorov, David and Zhukov, Dmitry and Barinova, Olga and Anton, Konushin and Vorontsova, Anna},
    booktitle = {2019 16th International Conference on Machine Vision Applications (MVA)},
    doi = {10.23919/mva.2019.8758020},
    language = {english},
    pages = {1--6},
    publisher = {IEEE},
    series = {IEEE Xplore Digital Library},
    title = {Measuring robustness of Visual SLAM},
    year = {2019}
}

Abstract:

Human gait is an important biometric index that allows to identify a person at a great distance without direct contact. Due to these qualities, which other popular identifiers such as fingerprints or iris do not have, the recognition of a person by the manner of walking has become very common in various areas where video surveillance systems can be used. With the development of computer vision techniques, a variety of approaches for human identification by movements in a video appear. These approaches are based both on natural biometric characteristics (human skeleton, silhouette, and their change during walking) and abstract features trained automatically which do not have physical justification. Modern methods combine classical algorithms of video and image analysis and new approaches that show excellent results in related tasks of computer vision, such as human identification by face and appearance or action and gesture recognition. However, due to the large number of conditions that can affect the walking manner of a person itself and its representation in video, the problem of identifying a person by gait still does not have a sufficiently accurate solution. Many methods are overfitted by the conditions presented in the databases on which they are trained, which limits their applicability in real life. In this paper, we provide a survey of state-of-the-art methods of gait recognition, their analysis and comparison on several popular video collections and for different formulations of the problem of recognition. We additionally reveal the problems that prevent the final solution of gait identification challenge.

Bibtex:

@article{sokolova19methods,
    address = {Road Town, United Kingdom},
    author = {Sokolova, A. and Konushin, A.},
    doi = {10.1134/s0361768819040091},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {4},
    pages = {213--220},
    publisher = {Road Town, United Kingdom},
    title = {Methods of Gait Recognition in Video},
    volume = {45},
    year = {2019}
}

Abstract:

We present a novel method for image anomaly detection, where algorithms that use samples drawn from some distribution of "normal" data, aim to detect out-of-distribution (abnormal) samples. Our approach includes a combination of encoder and generator for mapping an image distribution to a predefined latent distribution and vice versa. It leverages Generative Adversarial Networks to learn these data distributions and uses perceptual loss for the detection of image abnormality. To accomplish this goal, we introduce a new similarity metric, which expresses the perceived similarity between images and is robust to changes in image contrast. Secondly, we introduce a novel approach for the selection of weights of a multi-objective loss function (image reconstruction and distribution mapping) in the absence of a validation dataset for hyperparameter tuning. After training, our model measures the abnormality of the input image as the perceptual dissimilarity between it and the closest generated image of the modeled data distribution. The proposed approach is extensively evaluated on several publicly available image benchmarks and achieves state-of-the-art performance.

Bibtex:

@inproceedings{tuluptceva19perceptual,
    author = {Tuluptceva, Nina and Bakker, Bart and Fedulova, Irina and Konushin, Anton},
    booktitle = {Palaiahnakote S., Sanniti di Baja G., Wang L., Yan W. (eds) Pattern Recognition. ACPR 2019},
    doi = {10.1007/978-3-030-41404-7\_12},
    language = {english},
    pages = {164--178},
    publisher = {Springer},
    series = {Lecture Notes in Computer Science},
    title = {Perceptual Image Anomaly Detection},
    volume = {12046},
    year = {2019}
}

Bibtex:

@article{sokolova19posebased,
    author = {Sokolova, Anna and Konushin, Anton},
    doi = {10.1049/iet-bmt.2018.5046},
    issn = {2047-4938},
    journal = {IET BIOMETRICS},
    language = {english},
    number = {2},
    pages = {134--143},
    title = {Pose-based Deep Gait Recognition},
    volume = {8},
    year = {2019}
}

Abstract:

Modern computer vision methods usually require lots of labelled data for training. Besides price of labelling, problems with rare object classes and adaptation to new domain or task arise. One of the promising methods to solve these problems is to generate synthetic training data. In this work we focus on task of traffic sign detection. We consider several methods for generating synthetic data for training traffic sign detectors: random placement of signs of different quality (simple synthetic, CGI based and CGI improved using generative adversarial network). We also propose a method to replace real signs with synthetic signs. Experimental evaluation shows that proposed method improves quality of detection of rare traffic signs and that usage of synthetic data is very helpful for improving training of traffic sign classifier.

Projects:

Bibtex:

@inproceedings{shakhuro19traffic,
    author = {Shakhuro, Vlad and Faizov, Boris and Konushin, Anton},
    booktitle = {ICVIP 2019: Proceedings of the 3rd International Conference on Video and Image Processing},
    doi = {10.1145/3376067.3376105},
    language = {english},
    pages = {23--26},
    title = {Rare traffic sign recognition using synthetic training data},
    year = {2019}
}

Abstract:

Human gait is one of the biometric characteristics that a person can be identified by. However, the wide applicability of gait recognition in real life is prevented by a great variety of conditions that affect the gait representation, such as different viewpoints. In this work, we present a novel view resistant approach to overcome the multi-view recognition challenge. The new loss function is proposed to increase the stability of the model to view changes. Besides this, the cross-view embedding of the gait features is made to enhance their discriminant ability which improves the recognition accuracy as well. The proposed approaches show a significant gain in quality and allow to achieve the state-of-the-art accuracy on the most common benchmark and outperform the most successful model on the majority of the views and on average.

Bibtex:

@inproceedings{sokolova19resistant,
    author = {Sokolova, Anna and Konushin, Anton},
    booktitle = {ICVIP 2019: Proceedings of the 3rd International Conference on Video and Image Processing},
    doi = {10.1145/3376067.3376083},
    language = {english},
    pages = {7--12},
    title = {View resistant gait recognition},
    year = {2019}
}

Abstract:

Походка – важный биометрический показатель, позволяющий идентифицировать человека на большом расстоянии и без непосредственного контакта. Благодаря этим качествам, отсутствующим у других популярных идентификаторов, таких как отпечатки пальцев и радужная оболочка глаза, распознавание человека по походке в наши дни стало очень распространено и востребовано в различных сферах, где возможно использование систем видеонаблюдения. С развитием методов компьютерного зрения появляется множество подходов к идентификации человека по движениям в видео, использующих как естественные биометрические характеристики (скелет человека, его силуэт, их изменение во время ходьбы), так и абстрактные признаки. Современные методы объединяют в себе классические алгоритмы анализа видео и изображений и новые подходы, показывающие высокие результаты в смежных задачах компьютерного зрения, таких как идентификация человека по лицу или распознавание действий. Однако из-за большого количества условий, влияющих на саму манеру движения человека и ее представление в видео, задача идентификации человека по походке до сих пор не имеет достаточно точного решения. Многие методы заточены исключительно под условия, присутствующие в базах данных, на которых они обучаются, что ограничивает их применимость в реальной жизни. В данной работе проводится обзор современных методов распознавания человека по походке, их анализ и сравнение на нескольких популярных видео коллекциях и для разных формулировок задачи распознавания, а также выявляются проблемы, препятствующие окончательному решению задачи идентификации по походке.

Bibtex:

@article{sokolova19metody,
    author = {Соколова, А. И. and Конушин, А. С.},
    doi = {10.15514/ISPRAS-2019-31(1)-5},
    issn = {2220-6426; 2079-8156},
    journal = {Труды Института системного программирования РАН (электронный журнал)},
    language = {russian},
    number = {1},
    pages = {69--82},
    title = {Методы идентификации человека по походке в видео},
    volume = {31},
    year = {2019}
}

Abstract:

В работе рассматривается задача сопровождения людей в видеопотоке с целью их подсчета. Современные системы видеонаблюдения такие, как система видеонаблюдения Москвы, содержат сотни тысяч камер. Использование современных методов, разрабатываемых для работына одном компьютере с дорогим графическим ускорителем, не является экономически целесообразным для систем подобного масштаба. В работе предлагается распределенный алгоритм сопровождения. Он позволяет за счет детекции на разреженном множестве кадров сократить число необходимых вычислительных ресурсов. Детекция производится на серверах в центре обработки данных, а видеопоток обрабатывается локальными для камер узлами. Представлена методика оценки качества подсчета людей. Экспериментальная оценка показала, что предложенный алгоритм позволяет достичь приемлемого качества при частоте детекции 4/3 Гц.

Bibtex:

@article{kupljakov19raspredelennyj,
    author = {Купляков, Д. А. and Шальнов, Е. В. and Конушин, В. С. and Конушин, А. С.},
    doi = {10.1134/s0132347419040046},
    issn = {0132-3474},
    journal = {Программирование},
    language = {russian},
    number = {4},
    pages = {28--35},
    title = {Распределенный алгоритм сопровождения для подсчета людей в видео},
    year = {2019}
}

Bibtex:

@inproceedings{zvezdakova19barriers,
    author = {Zvezdakova, Anastasia and Kulikov, Dmitriy and Kondranin, Denis and Vatolin, Dmitriy},
    booktitle = {GraphiCon'2019 Proceedings},
    doi = {10.30987/graphicon-2019-2-22-27},
    language = {english},
    pages = {22--27},
    publisher = {CEUR Workshop Proceedings},
    title = {Barriers Towards No-reference Metrics Application to Compressed Video Quality Analysis: on the Example of No-reference Metric NIQE},
    volume = {2485},
    year = {2019}
}

Bibtex:

@inproceedings{zvezdakova19hacking,
    author = {Zvezdakova, Anastasia and Zvezdakov, Sergey and Kulikov, Dmitriy and Vatolin, Dmitriy},
    booktitle = {GraphiCon'2019 Proceedings},
    doi = {10.30987/graphicon-2019-2-53-57},
    language = {english},
    pages = {53--57},
    publisher = {CEUR Workshop Proceedings},
    title = {Hacking VMAF with Video Color and Contrast Distortion},
    volume = {2485},
    year = {2019}
}

Abstract:

Recent advances in deep learning have markedly improved the quality of visual-attention modelling. In this work we apply these advances to video compression. We propose a compression method that uses a saliency model to adaptively compress frame areas in accordance with their predicted saliency. We selected three state-of-the-art saliency models, adapted them for video compression and analyzed their results. The analysis includes objective evaluation of the models as well as objective and subjective evaluation of the compressed videos. Our method, which is based on the x264 video codec, can produce videos with the same visual quality as regular x264, but it reduces the bitrate by 25\% according to the objective evaluation and by 17\% according to the subjective one. Also, both the subjective and objective evaluations demonstrate that saliency models can compete with gaze maps for a single observer. Our method can extend to most video bitstream formats and can improve video compression quality without requiring a switch to a new video encoding standard.

Bibtex:

@inproceedings{lyudvichenko19improving,
    address = {New York, N.Y., United States},
    author = {Lyudvichenko, Vitaliy and Erofeev, Mikhail and Ploshkin, Alexander and Vatolin, Dmitriy},
    booktitle = {Proceedings of the 2019 International Conference on Intelligent Medicine and Image Processing},
    doi = {10.1145/3332340.3332358},
    language = {english},
    pages = {88--94},
    publisher = {New York, N.Y., United States},
    series = {IMIP '19},
    title = {Improving Video Compression With Deep Visual-Attention Models},
    year = {2019}
}

Abstract:

The field of automatic image inpainting has progressed rapidly in recent years, but no one has yet proposed a standard method of evaluating algorithms. This absence is due to the problem's challenging nature: image-inpainting algorithms strive for realism in the resulting images, but realism is a subjective concept intrinsic to human perception. Existing objective image-quality metrics provide a poor approximation of what humans consider more or less realistic. To improve the situation and to better organize both prior and future research in this field, we conducted a subjective comparison of nine state-of-the-art inpainting algorithms and propose objective quality metrics that exhibit high correlation with the results of our comparison.

Bibtex:

@inproceedings{molodetskikh19perceptually,
    author = {Molodetskikh, Ivan and Erofeev, Mikhail and Vatolin, Dmitriy},
    booktitle = {Proceedings of the 29th International Conference on Computer Graphics and Vision},
    doi = {10.30987/graphicon-2019-2-131-135},
    language = {english},
    pages = {131--135},
    publisher = {CEUR Workshop Proceedings},
    title = {Perceptually Motivated Method for Image Inpainting Comparison},
    volume = {2485},
    year = {2019}
}

Abstract:

This paper presents a new way of getting high-quality saliency maps for video, using a cheaper alternative to eye-tracking data. We designed a mouse-contingent video viewing system which simulates the viewers' peripheral vision based on the position of the mouse cursor. The system enables the use of mouse-tracking data recorded from an ordinary computer mouse as an alternative to real gaze fixations recorded by a more expensive eye-tracker. We developed a crowdsourcing system that enables the collection of such mouse-tracking data at large scale. Using the collected mouse-tracking data we showed that it can serve as an approximation of eye-tracking data. Moreover, trying to increase the efficiency of collected mouse-tracking data we proposed a novel deep neural network algorithm that improves the quality of mouse-tracking saliency maps.

Bibtex:

@inproceedings{lyudvichenko19predicting,
    author = {Lyudvichenko, Vitaliy and Vatolin, Dmitriy},
    booktitle = {Proceedings of the 29th International Conference on Computer Graphics and Vision},
    doi = {10.30987/graphicon-2019-2-127-130},
    language = {english},
    pages = {127--130},
    publisher = {CEUR Workshop Proceedings},
    title = {Predicting video saliency using crowdsourced mouse-tracking data},
    volume = {2485},
    year = {2019}
}

Abstract:

В данной работе рассматривается задача построения минимального репрезентативного набора физических признаков видео, которые могут быть использованы в разных регрессионных и классификационных задачах при работе с видео. Для тестирования полученного набора признаков была рассмотрена задача классификации видео на 4 класса: мультфильм, съемка с дрона, компьютерная игра и спортивная трансляция. В работе описан автоматический подход для создания размеченной по жанрам базы видеопоследовательностей, его аугментации и очистки от аномальных данных. С помощью этого подхода для проведения экспериментальной оценки была создана база из 14271 видео, где минимальное количество видео на класс составляет 2700 видео. Методом градиентного бустинга был обучен ансамбль решающих деревьев, средняя точность и полнота прогнозирования которого на тестовой выборке составили 86.15\% и 86.12\%. Также были опробованы другие методы машинного обучения такие, как логистическая регрессия, наивный гауссовский классификатор, метод опорных векторов и случайный лес, и они показали более худшие результаты. Наиболее важными признаками для принятия решений в задаче классификации видео оказались две метрики детализации изображения, вычисляемые с помощью оператора Лапласа и на основе эффекта повторного размытия, и пространственная сложность.

Bibtex:

@article{kazantsev19primenenie,
    address = {Москва},
    author = {Казанцев, Р. А. and Звездаков, С. В. and Ватолин, Д. С.},
    doi = {10.24412/FgiBa7Wglt4},
    issn = {2307-8162},
    journal = {International Journal of Open Information Technologies},
    language = {russian},
    number = {5},
    pages = {34--38},
    publisher = {Москва},
    title = {Применение физических признаков видео к задаче классификации},
    volume = {7},
    year = {2019}
}

Bibtex:

@inproceedings{shcherbakov19dynamic,
    author = {Shcherbakov, A. S. and Frolov, V. A.},
    booktitle = {WSCG'2019 Full papers proceedings},
    language = {english},
    pages = {83--90},
    publisher = {Computer Science Research Notes 2901 Pilsen, Czech Republic},
    series = {2901},
    title = {Dynamic Radiosity},
    year = {2019}
}

Abstract:

Hardware acceleration of ray tracing is an active research field, but only with the release of Nvidia Turing architecture GPUs it became widely available. Nvidia RTX is a proprietary hardware ray tracing acceleration technology available in Vulkan and DirectX APIs as well as through Nvidia OptiX. Since the implementation details are unknown to the public, there are a lot of questions about what it actually does under the hood. To find answers to these questions, we implemented classic path tracing algorithm using RTX via both DirectX and Vulkan and conducted several experiments with it to investigate the inner workings of this technology. We tested actual hardware implementation of RTX technology on RTX2070 GPU and the software fallback in the driver on GTX1070 GPU. In this paper we present results of these experiments and speculate on the internal architecture of RTX.

Bibtex:

@article{sanzharov19examination,
    address = {Aachen, Germany},
    author = {Sanzharov, V. V. and Gorbonosov, A. I. and Frolov, V. A. and Voloboy, A. G.},
    doi = {10.30987/graphicon-2019-2-7-12 http://ceur-ws.org/Vol-2485/paper3.pdf},
    issn = {1613-0073},
    journal = {CEUR Workshop Proceedings},
    language = {english},
    pages = {7--12},
    publisher = {Aachen, Germany},
    title = {Examination of the Nvidia RTX},
    volume = {2485},
    year = {2019}
}

Bibtex:

@article{shcherbakov19implementation,
    address = {Aachen, Germany},
    author = {Shcherbakov, A. S. and Frolov, V. A.},
    doi = {10.30987/graphicon-2019-2-13-17 http://ceur-ws.org/Vol-2485/paper4.pdf},
    issn = {1613-0073},
    journal = {CEUR Workshop Proceedings},
    language = {english},
    pages = {13--17},
    publisher = {Aachen, Germany},
    title = {Implementation of the Radiosity algorithm for large scale scenes},
    volume = {2485},
    year = {2019}
}

Bibtex:

@article{sanzharov19restricted,
    address = {Aachen, Germany},
    author = {Sanzharov, V. V. and Frolov, V. A. and Pavlov, I. V.},
    doi = {10.30987/graphicon-2019-2-37-42 http://ceur-ws.org/Vol-2485/paper9.pdf},
    issn = {1613-0073},
    journal = {CEUR Workshop Proceedings},
    language = {english},
    pages = {37--42},
    publisher = {Aachen, Germany},
    title = {Restricted extensions for GPU photo-realistic renderer},
    volume = {2485},
    year = {2019}
}

Abstract:

Сегодня имеется более 200 различных программных продуктов для фотореалистичного синтеза изображений, но отсутствуют методы их объективного сравнения, что не позволяет выбрать наилучшую систему для каждого конкретного случая. Чтобы показать преимущества именно своей системы, разработчики подбирают сцены, позволяющие в наилучшем свете показать превосходство используемых в их продукте методов. Однако такие ухищрения не гарантируют, что предложенный метод будет хорошо работать на других сценах. Для объективной оценки рендер-систем требуется открытый бенчмарк тестовых сцен. http://www.osp.ru/os/2019/02/13054955/

Bibtex:

@article{frolov19vybor,
    author = {Фролов, Владимир and Галактионов, Владимир and Павлов, Денис and Трофимов, Максим and Казбеев, Павел},
    journal = {Открытые системы},
    language = {russian},
    number = {2},
    pages = {30--33},
    title = {Выбор системы рендеринга: Open Rendering Benchmark},
    year = {2019}
}

Bibtex:

@inproceedings{frolov19generatsija,
    address = {Москва},
    author = {Фролов, В. А. and Павлов, Д. С.},
    booktitle = {Научная конференция ЛОМОНОСОВСКИЕ ЧТЕНИЯ. Тезисы докладов. 15-25 апреля 2019 г.},
    language = {russian},
    pages = {86--87},
    publisher = {Москва},
    series = {СЕКЦИЯ ВЫЧИСЛИТЕЛЬНОЙ МАТЕМАТИКИ И КИБЕРНЕТИКИ, Факультет вычислительной математики и кибернетики},
    title = {Генерация синтетических данных движений человека для приложений машинного обучения},
    year = {2019}
}

Abstract:

Comparison of algorithms and methods is the cornerstone of any research which allows to prove that the proposed approach takes a step forward with a previous one. In areas such as computer vision or machine learning a huge amount of open data sets have been created, in which researchers can test their algorithms and compete with each other. In computer graphics and photorealistic rendering at the current moment the situation is different. Open sets of scenes in which different researchers using different render systems could get matching (or at least close) images do not exist. This leads to the fact that in scientific papers so-called “Cherry-picking” is practiced. This one is a pedantic selection of scenes and lighting conditions in such a way as to demonstrate the advantages of the developed algorithm. Such approach greatly reduces the practical significance of research - even if the new method works well on a certain class of scenes, it does not mean it will work for other cases. For this reason, many render systems still do not go beyond the basic algorithms, not trusting the results of research. In this paper, we began to fill this gap. We have created a special set of scenes (the so-called “benchmark”), which allows us to evaluate the performance of lighting integration in various situations and, thus, to show the positive and negative aspects of render systems and algorithms used by them in various conditions. We recreated many scenes from well-known computer graphics papers and added some scenes based on our experience. Our goal is to make the most complete coverage possible, using as few scenes and functionality of the render system as possible so that such a comparison can be easily reproduced in any existing system. To validate our approach, we conducted a pilot comparison among 4 popular products for 3D Studio Max (VRay, Corona, Octane, Hydra Renderer) for the speed of integrating the lighting on various scenarios. Despite the fact that 3 out of 4 systems are closed commercial products, we managed to get the same or similar images for all scenes, which indicates the viability of our proposed approach. https://lppm3.ru/files/journal/XLV/MathMontXLV-Frolov.pdf

Bibtex:

@article{frolov19otkrytoe,
    author = {Фролов, В. and Павлов, Д. and Трофимов, М. and Казбеев, П. and Галактионов, В.},
    doi = {10.20948/mathmontis-2019-45-9},
    journal = {Matematica Montisnigri},
    language = {russian},
    pages = {106--122},
    title = {Открытое соревнование рендер-систем},
    year = {2019}
}

Abstract:

Сравнительный анализметодов синтеза текстурпо образцуА.Ю.Бабичев1,В.А.Фролов1,2andrey.babichev@graphics.cs.msu.ru|vfrolov@graphics.cs.msu.ru.1Московский Государственный Университет имени М.В. Ломоносова, Москва, Россия;2Институт прикладной математики имени М.В, Келдыша РАН.В даннойстатье предлагается сравнительный анализ существующих методовсинтеза текстур. Выявляются преимущества и недостаткисуществующихметодов.В нашей работе мы руководствовались результатами обширного экспертного тестирования, в котором приняло участие более 25человек на 20 принципиально разных текстурах. Это позволилонамсделать выводы о том, на каких типах текстур какие методы следует использовать.Наше исследование показало, что популярные в последнее время нейро-сетевые и статистические методы не являются лучшими ни по качеству синтеза, ни по скорости. Однако при этом они генерируют более разнообразные текстуры. С другой стороны, наиболее простой и быстрый метод переставления патчей показал лучшее качество и скорость. Таким образом, синтез текстур по образцу –одна из перспективных областей исследования, в которой различные подходы обладают разными преимуществами.

Bibtex:

@inproceedings{babichev19sravnitelnyj,
    author = {Бабичев, А. Ю. and Фролов, В. А.},
    booktitle = {Труды Международной конференции по компьютерной графике и зрению "Графикон", Брянск, Россия, 23-26 сентября},
    doi = {10.30987/graphicon-2019-1-21-25 https://www.graphicon.ru/html/2019/papers/01/Volume1\_paper\_01.pdf},
    language = {russian},
    publisher = {https://www.graphicon.ru/html/2019/papers/01/Volume1\_paper\_01.pdf},
    series = {29},
    title = {Сравнительный анализ методов синтеза текстур по образцу},
    volume = {2019},
    year = {2019}
}

Bibtex:

@article{sanzharov19uroven,
    author = {Санжаров, В. В. and Фролов, В. А.},
    doi = {10.1134/S0132347419040071},
    issn = {0132-3474},
    journal = {Программирование},
    language = {russian},
    number = {4},
    pages = {54--63},
    title = {Уровень детализации для предрасчитанных процедурных текстур},
    year = {2019}
}

Projects:

Bibtex:

@inproceedings{senyukova19automated,
    author = {Senyukova, Olga V. and Brotikovskaya, Danuta S. and Gorokhova, Svetlana G. and Tebenkova, Ekaterina S.},
    booktitle = {Computational Intelligence, 9th International Joint Conference, IJCCI 2017 Funchal-Madeira, Portugal, November 1-3, 2017 Revised Selected Papers},
    doi = {10.1007/978-3-030-16469-0\_12},
    language = {english},
    pages = {221--238},
    publisher = {Springer, Cham Switzerland},
    series = {Studies in Computational Intelligence},
    title = {Automated Diagnostic Model Based on Isoline Map Analysis of Myocardial Tissue Structure},
    volume = {829},
    year = {2019}
}

Projects:

Bibtex:

@article{gavrishchaka19synergy,
    author = {Gavrishchaka, Valeriy and Senyukova, Olga and Koepke, Mark},
    doi = {10.1080/23746149.2019.1582361},
    issn = {2374-6149},
    journal = {ADVANCES IN PHYSICS-X},
    language = {english},
    number = {1},
    pages = {1582361},
    title = {Synergy of physics-based reasoning and machine learning in biomedical applications: towards unlimited deep learning with limited data},
    volume = {4},
    year = {2019}
}

Abstract:

The authors consider the problem of human pose estimation using probabilistic convolutional neural networks. They explore ways to improve human pose estimation accuracy on standard pose estimation benchmarks MPII human pose and Leeds Sports Pose (LSP) datasets using frameworks for probabilistic deep learning. Such frameworks transform deterministic neural network into a probabilistic one and allow sampling of independent and equiprobable hypotheses (different outputs) for a given input. Overlapping body parts and body joints hidden under clothes or other obstacles make the problem of human pose estimation ambiguous. In this context to get accurate estimation of joints’ position they use uncertainty in network's predictions, which is represented by variance of hypotheses, provided by a probabilistic convolutional neural network, and confidence is characterised by mean of them. Their work is based on current CNN cascades for pose estimation. They propose and evaluate three probabilistic convolutional neural networks built on top of deterministic ones with two probabilistic deep learning frameworks – DISCO networks and Bayesian SegNet. The authors evaluate their models on standard pose estimation benchmarks and show that proposed probabilistic models outperform base deterministic ones.

Bibtex:

@article{ilia18probabilistic,
    address = {United Kingdom},
    author = {Ilia, Petrov and Shakhuro, Vlad and Konushin, Anton},
    doi = {10.1049/iet-cvi.2017.0382},
    issn = {1751-9640; 1751-9632},
    journal = {IET Computer Vision},
    language = {english},
    number = {5},
    pages = {578--585},
    publisher = {United Kingdom},
    title = {Deep probabilistic human pose estimation},
    volume = {12},
    year = {2018}
}

Abstract:

This work is focused on person identification task in video sequences. For this task we propose two complementing solutions, which can be applied in different cases: gait and visual recognition. For gait recognition three kinds of features are used: anthropometric features, based on the length of the skeleton segments; relative distance features, based on relative distances between the skeleton joints; and motion features, based on the movement of a joint between two frames. Two versions of the gait recognition algorithm are presented: the first one uses the depth data alongside with the images while the other one uses only the video sequence. For visual recognition from appearance we propose a deep learning algorithm that returns binary image features. Each algorithm was tested on two datasets. Furthermore, we perform experiments on transfer from one dataset to another to check trained model transferability.

Bibtex:

@article{arseev18human,
    address = {Road Town, United Kingdom},
    author = {Arseev, Sergey and Konushin, Anton and Liutov, Vladimir},
    doi = {10.1134/S0361768818040035},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {4},
    pages = {258--265},
    publisher = {Road Town, United Kingdom},
    title = {Human Recognition by Appearance and Gait},
    volume = {44},
    year = {2018}
}

Abstract:

Походка – важный биометрический показатель, позволяющий идентифицировать человека без личного контакта. В данной работе проводится обзор современных методов распознавания человека по походке, их анализ и сравнение, а также выявляются проблемы, препятствующие окончательному решению задачи распознавания по походке.

Bibtex:

@inproceedings{sokolova18obzor,
    author = {Соколова, А. И. and Конушин, А. С.},
    booktitle = {Proceedings of the 28th International Conference on Computer Graphics and Vision GraphiCon'2018},
    language = {russian},
    pages = {234--237},
    publisher = {Tomsk Polytechnic University Tomsk, Russia},
    title = {Обзор методов распознавания человека по походке в видео},
    year = {2018}
}

Abstract:

В работе рассматривается задача сопровождения людей в видеопотоке с целью их подсчёта. Современные системы видеонаблюдения такие, как система видеонаблюдения Москвы, содержат сотни тысяч камер. Использование современных методов, разрабатываемых для работы на одном компьютере с дорогим графическим ускорителем, не является экономически целесообразным для систем подобного масштаба. В работе предлагается распределённый алгоритм сопровождения. Он позволяет за счёт детекции на разреженном множестве кадров сократить число необходимых вычислительных ресурсов. Детекция производится на серверах в центре обработки данных, а видеопоток обрабатывается локальными для камер узлами. Представлена методика оценки качества подсчёта людей. Экспериментальная оценка показала, что предложенный алгоритм позволяет достичь приемлемого качества при частоте детекции 4/3 Гц.

Bibtex:

@inproceedings{kupljakov18raspredelennyj,
    author = {Купляков, Д. А. and Шальнов, Е. В. and Конушин, В. С. and Конушин, А. С.},
    booktitle = {GraphiCon 2018: труды 28-й Междунар. конф. по компьютерной графике и машинному зрению},
    language = {russian},
    pages = {208--213},
    publisher = {Нац. исслед. Том. политех. ун-т Томск},
    series = {GraphiCon},
    title = {Распределённый алгоритм сопровождения людей в видео},
    year = {2018}
}

Abstract:

В работе рассматривается метод генерации синтетических обучающих выборок для задачи классификации дорожных знаков. Метод основан на использовании порождающих конкурирующих нейросетей и метрики Васерштейна. Исследуется метод условной генерации изображений, когда на вход порождающей нейросети подается случайный шум и метка класса изображения, которое нужно сгенерировать. Для обучения такой нейросети предлагается использовать перекрестную энтропию в добавление к метрике Васерштейна. Для стабилизации процесса обучения используются веса для обучающей выборки. Экспериментальная оценка метода показывает, что условная порождающая сеть работает лучше, чем простая генерация дорожных знаков по иконке, однако не дотягивает до метода, в котором для каждого класса обучается отдельная порождающая нейросеть

Projects:

Bibtex:

@inproceedings{hrushkov18sintez,
    author = {Хрушков, П. В. and Шахуро, В. И. and Конушин, А. С.},
    booktitle = {Proceedings of the 28th International Conference on Computer Graphics and Vision GraphiCon'2018},
    language = {russian},
    pages = {242--246},
    publisher = {Tomsk Polytechnic University Tomsk, Russia},
    title = {Синтез изображений дорожных знаков с помощью условных порождающих противоборствующих нейросетей},
    year = {2018}
}

Abstract:

В работе исследуется применимость порождающих конкурирующих нейронных сетей для синтеза обучающих выборок на примере задачи классификации дорожных знаков. Рассматриваются порождающие нейронные сети, обучаемые с помощью метрики Васерштейна. В качестве базового метода для сравнения используется метод генерации синтетических изображений дорожных знаков по иконке. Проводится экспериментальное сравнение нейросетевых классификаторов, обученных на реальных данных, двух видах синтетических данных, а также смеси реальных и синтетических данных. Эксперименты показывают, что современные порождающие нейронные сети позволяют создавать реалистичные обучающие выборки для классификации автодорожных знаков, которые превосходят по качеству методы генерации знаков по иконкам, но немного уступают по качеству реальным данным.

Projects:

Bibtex:

@article{shahuro18sintez,
    author = {Шахуро, В. И. and Конушин, А. С.},
    doi = {10.18287/2412-6179-2018-42-1-105-112},
    issn = {0134-2452},
    journal = {Компьютерная оптика},
    language = {russian},
    number = {1},
    pages = {105--112},
    title = {Синтез обучающих выборок для  классификации дорожных знаков с помощью нейросетей},
    volume = {42},
    year = {2018}
}

Abstract:

Despite the higher video-completion quality that recently proposed methods have enabled for a wide variety of cases, their computational complexity remains a major concern. These methods typically frame video completion as an optimization problem over the whole spatiotemporal domain-a problem that is expensive to solve both in time and space. In this paper we propose a lighter-weight multipass video-completion pipeline that replaces global spatiotemporal optimization with simpler frame-by-frame motion reconstruction and refinement. We achieve a processing speed of 2.6 seconds per frame on Full HD content while delivering nearly state-of-the-art completion quality for a wide range of dynamic scenes captured using a free-moving camera. To validate the performance of our proposed method, we conducted a subjective comparison of different video-completion results for 26 test sequences from the DAVIS data set.

Bibtex:

@inproceedings{bokov18times,
    author = {Bokov, A. and Vatolin, D.},
    booktitle = {2018 25th IEEE International Conference on Image Processing (ICIP)},
    doi = {10.1109/ICIP.2018.8451683},
    language = {english},
    pages = {2122--2126},
    publisher = {Athens, Greece},
    title = {100+ Times Faster Video Completion by Optical-Flow-Guided Variational Refinement},
    year = {2018}
}

Abstract:

Video synchronization is a fundamental computer-vision task that is necessary for a wide range of applications. A 3D video involves two streams, which show the scene from different angles concurrently, but many cases exhibit desynchronization between them. This paper investigates the problem of synchronizing the left and right stereoscopic views. We assume the temporal shift (time difference) and geometric distortion between the two streams are constant throughout each scene. We propose a temporal-shift estimation method with subframe accuracy based on a block-matching algorithm.

Bibtex:

@inproceedings{ploshkin18accurate,
    author = {Ploshkin, Aleksandr and Vatolin, Dmitriy},
    booktitle = {Proceedings of the 2018 3DTV Conference: The True Vision - Capture, Transmission and Display of 3D Video (3DTV-CON)},
    doi = {10.1109/3DTV.2018.8478431},
    language = {english},
    pages = {1--4},
    title = {Accurate Method of Temporal Shift Estimation for 3D Video},
    year = {2018}
}

Abstract:

Channel mismatch (the result of swapping left and right views) is a 3D-video artifact that can cause major viewer discomfort. This work presents a novel high-accuracy method of channel-mismatch detection. In addition to the features described in our previous work, we introduce a new feature based on a convolutional neural network; it predicts channel-mismatch probability on the basis of the stereoscopic views and corresponding disparity maps. A logistic-regression model trained on the described features makes the final prediction. We tested this model on a set of 900 stereoscopic-video scenes, and it outperformed existing channel-mismatch detection methods that previously served in analyses of full-length stereoscopic movies.

Bibtex:

@inproceedings{lavrushkin18channelmismatch,
    author = {Lavrushkin, Sergey and Vatolin, Dmitriy},
    booktitle = {Proceedings of the 2018 3DTV Conference: The True Vision - Capture, Transmission and Display of 3D Video (3DTV-CON)},
    doi = {10.1109/3DTV.2018.8478542},
    language = {english},
    pages = {1--4},
    title = {Channel-Mismatch Detection Algorithm for Stereoscopic Video Using Convolutional Neural Network},
    year = {2018}
}

Abstract:

Occlusion filling is a basic problem for multiview video generation from existing monocular video. The essential goal of this problem is to recover missing information about a scenes 3D structure and corresponding texture.We propose a method for content-aware deformation of the source view that ensures no disoccluded regions are visible in the synthesized views while also keeping visible distortions to a minimum. We formulate this problem in terms of global energy min-imization. Furthermore, we introduce a similar variable-rejection algorithm that, along with other known optimization techniques, allows us to accelerate the energy function minimization by nearly 30 times and still maintain the visual quality of the synthesized views.

Bibtex:

@inproceedings{khatiullin18occlusion,
    author = {Khatiullin, Aidar and Erofeev, Mikhail and Vatolin, Dmitriy},
    booktitle = {Proceedings of the 2018 3DTV Conference: The True Vision - Capture, Transmission and Display of 3D Video (3DTV-CON)},
    doi = {10.1109/3DTV.2018.8478562},
    language = {english},
    pages = {1--4},
    title = {Fast Occlusion Filling Method For Multiview Video Generation},
    year = {2018}
}

Abstract:

Many factors can cause color distortions between stereoscopic views during 3D-video shooting. Numerous viewers experience discomfort and headaches when watching stereoscopic videos that contain such distortions. In addition, 3D videos with color differences are hard to process because many algorithms assume brightness constancy.We propose an automatic method for correcting color distortions between stereoscopic views and compare it with analogs. The comparison shows that our proposed method combines high color-correction accuracy with relatively low computational complexity.

Bibtex:

@inproceedings{lavrushkin18local,
    author = {Lavrushkin, Sergey and Lyudvichenko, Vitaliy and Vatolin, Dmitriy},
    booktitle = {Proceedings of the 2018 3DTV Conference: The True Vision - Capture, Transmission and Display of 3D Video (3DTV-CON)},
    doi = {10.1109/3DTV.2018.8478453},
    language = {english},
    pages = {1--4},
    title = {Local Method of Color-Difference Correction Between Stereoscopic-Video Views},
    year = {2018}
}

Abstract:

Recent advances in video completion have made it possible to plausibly fill in large missing regions for a wide variety of unconstrained scenes captured with a freely moving camera. These advances enable automatic generation of high-quality background plates for filling holes during virtual-view synthesis. Most hole-filling approaches, however, still tend to impose restrictive assumptions, such as approximating the background as a single plane, or they forgo background reconstruction by copying from other input-video frames altogether. Moreover, all existing approaches either directly or indirectly rely on binary separation of the scene into background and foreground, which is sometimes not possible. In this paper we propose a multilayer extension of the video-completion algorithm in [5] to RGBD video and demonstrate its advantages compared with alternative approaches when filling holes for 3D-view synthesis.

Bibtex:

@inproceedings{bokov18multilayer,
    author = {Bokov, Alexander and Vatolin, Dmitriy},
    booktitle = {2018 International Conference on 3D Imaging (IC3D)},
    doi = {10.1109/IC3D.2018.8657873},
    language = {english},
    pages = {1--8},
    publisher = {Brussels},
    title = {Multilayer RGBD-Video Completion For Hole Filling in 3D-View Synthesis},
    year = {2018}
}

Abstract:

Video-completion methods aim to complete selected regions of a video sequence in a natural looking manner with little to no additional user interaction. Numerous algorithms were proposed to solve this problem; however, a unified benchmark to quantify the progress in the field is still lacking. Video-completion results are usually judged by their plausibility and aren’t expected to adhere to one ground-truth result, which complicates measuring the video-completion performance. In this paper, we address this problem by proposing a set of full-reference quality metrics that outperform naïve approaches and an online benchmark for video-completion algorithms. We construct seven test sequences with ground-truth video-completion results by composing various foreground objects over a set of background videos. Using this dataset, we conduct an extensive comparative study of video-completion perceptual quality involving six algorithms and over 300 human participants. Finally, we show that by relaxing the requirement of complete adherence to ground truth and by taking into account temporal consistency we can increase the correlation of objective quality metrics with perceptual completion quality on the proposed dataset.

Bibtex:

@article{bokov18toward,
    address = {Germany},
    author = {Bokov, Alexander and Vatolin, Dmitriy and Erofeev, Mikhail and Gitman, Yury},
    doi = {10.1007/s11760-018-1387-5},
    issn = {1863-1711; 1863-1703},
    journal = {Signal, Image and Video Processing},
    language = {english},
    pages = {1--8},
    publisher = {Germany},
    title = {Toward an objective benchmark for video completion},
    year = {2018}
}

Abstract:

Ежегодно выпускается большое количество фильмов в стереофор- мате. В процессе создания стереоконтента возникают специфические проблемы, отражающиеся на качестве получаемого продукта. Автомати- ческие системы анализа фильмов во многих случаях опираются на карты диспаратности, алгоритмы построения которых работают хуже при нали- чии определённых артефактов, что может привести к снижению точности анализа. В данной работе предлагается по стереопаре выбирать алгоритм, устойчивый к имеющимся на ней артефактам. При рассмотрении 28 мо- дификаций пяти алгоритмов точность выбора на вручную размеченном наборе данных составила 0,907.

Bibtex:

@inproceedings{muhamedov18avtomaticheskij,
    author = {Мухамедов, А. М. and Ватолин, Д. С. and Долганов, С. В.},
    booktitle = {Запись и воспроизведение объемных изображений в кинематографе и других областях: X Международная научно-практическая конференция},
    language = {russian},
    pages = {276--286},
    publisher = {ВГИК Москва},
    title = {Автоматический выбор алгоритма построения карт диспаратности для стереопар с артефактами съёмки или конвертации, устойчивого к данным артефактам},
    year = {2018}
}

Bibtex:

@article{molodetskih18algoritm,
    author = {Молодецких, И. А. and Ватолин, Д. С.},
    issn = {1991-3400},
    journal = {Мир техники кино},
    language = {russian},
    number = {2018-1(12)},
    pages = {18--22},
    title = {Алгоритм автоматического построения карт перемонтажа разных версий фильма},
    year = {2018}
}

Abstract:

Одним из распространенных способов создания стереофильмов является конвертация из 2D-формата в стереоформат. Это процесс, в ходе которого из одного изображения создаются два новых — для левого и правого глаза. Для заполнения возникающих областей открытия исполь- зуются различные приёмы: деформация заднего плана, масштабирование объектов, удаление незначительных для сюжета предметов. Все они тре- буют большого количества ручной работы, поэтому при некачественном выполнении могут образовываться дефекты, которые вызывают у зрите- лей ощущение дискомфорта и головную боль. Разработан метод автоматического поиска ограниченного набора приёмов конвертации видео из 2D-версии в стереоформат.

Bibtex:

@inproceedings{pereverzeva18metod,
    author = {Переверзева, П. А. and Ватолин, Д. С. and Хатиуллин, А. А.},
    booktitle = {Запись и воспроизведение объемных изображений в кинематографе и других областях: X Международная научно-практическая конференция},
    language = {russian},
    pages = {241--261},
    publisher = {ВГИК Москва},
    title = {Метод автоматического поиска различий между 2d и стереоскопическими версиями фильмов},
    year = {2018}
}

Abstract:

В настоящее время большое количество фильмов производится в стереоскопическом формате. Однако многие фильмы, выходящие в про- кат, содержат искажения, вызывающие дискомфорт при просмотре. Для контроля качества производимых стереоскопических видео существуют автоматические алгоритмы, оценки которых не отражают степень болез- ненности анализируемых искажений для зрителей. Данная работа посвя- щена созданию метода прогнозирования степени дискомфорта зрителей, который может возникнуть при просмотре стереофильмов, содержащих геометрические, цветовые и временные искажения ракурсов. Предло- женная модель была применена для анализа 60 полнометражных стере- офильмов, который показал, что даже среди низкобюджетных фильмов существуют фильмы, при просмотре которых зрители не почувствуют дискомфорта.

Bibtex:

@inproceedings{antsiferova18prognoziruemaja,
    author = {Анциферова, А. В. and Ватолин, Д. С.},
    booktitle = {Запись и воспроизведение объемных изображений в кинематографе и других областях: X Международная научно-практическая конференция},
    language = {russian},
    pages = {73--89},
    publisher = {ВГИК Москва},
    title = {Прогнозируемая оценка степени дискомфорта зрителей, вызываемого геометрическими, цветовыми и временными искажениями стереоскопического видео, для 60 снятых стереофильмов},
    year = {2018}
}

Abstract:

При стереосъёмке из-за ошибок в калибровке двух камер могут возникнуть геометрические искажения ракурсов стереофильма и несо- ответствие в них резкости. Поиск и анализ таких искажений в снятых стереофильмах производится автоматически с последующей визуальной оценкой заметности найденных искажений экспертами. В данной рабо- те описан алгоритм, позволяющий автоматизировать процесс выделения фрагментов стереокадра, содержащих наиболее заметные геометрические искажения ракурсов и несоответствие ракурсов по резкости.

Bibtex:

@inproceedings{kondranin18razrabotka,
    author = {Кондранин, Д. С. and Ватолин, Д. С.},
    booktitle = {Запись и воспроизведение объемных изображений в кинематографе и других областях: X Международная научно-практическая конференция},
    language = {russian},
    pages = {287--300},
    publisher = {ВГИК Москва},
    title = {Разработка метода выделения областей искажения на кадрах стереоскопического видео, содержащих геометрические искажения или расхождения по резкости},
    year = {2018}
}

Bibtex:

@article{lavrushkin18razrabotka,
    author = {Лаврушкин, С. В. and Ватолин, Д. С.},
    issn = {1991-3400},
    journal = {Мир техники кино},
    language = {russian},
    number = {2018-1(12)},
    pages = {12--17},
    title = {Разработка метода поиска перепутанных ракурсов в стереофильмах},
    year = {2018}
}

Abstract:

Статья посвящена работе по устранению недостатков существую- щего алгоритма оценки несоответствий по резкости. Была определена наиболее значимая проблема — ошибки второго рода на стереоскопиче- ских кадрах, содержащих источники света. Основная часть работы по- священа разработке метода определения присутствия источника света в стереопаре. Результаты работы разработанного метода подтверждаются экспериментами, проведёнными на полнометражных стереоскопических кинофильмах.

Bibtex:

@inproceedings{groholskij18sokraschenie,
    author = {Грохольский, С. Д. and Ватолин, Д. С.},
    booktitle = {Запись и воспроизведение объемных изображений в кинематографе и других областях: X Международная научно-практическая конференция},
    language = {russian},
    pages = {262--275},
    publisher = {ВГИК Москва},
    title = {Сокращение числа ложноположительных срабатываний в алгоритме оценки расхождения по резкости между ракурсами в видео},
    year = {2018}
}

Bibtex:

@inproceedings{vatolin18uvelichenie,
    author = {Ватолин, Д. С. and Ерофеев, М. В.},
    booktitle = {Ломоносовские чтения 2018 ф-т ВМК МГУ},
    language = {russian},
    pages = {56--57},
    publisher = {Макс-Пресс},
    title = {Увеличение эффективности субъективного удаленного сравнения видиопоследовательностей},
    year = {2018}
}

Abstract:

In this work we propose way of combining ordinary and Markov chain Monte Carlo rendering techniques in image space. We used per-pixel mask to separate pixels (which we want to run Markov chain on) from the rest of the image. The mask was obtained from the ordinary Monte Carlo noise analysis. The proposed method was tested with combination of Multiplexed Metropolis Light Transport as Markov chain technique and two ordinary Monte Carlo rendering techniques - Instant Bidirectional Path Tracing and Light Tracing. As a result, our method allows us to get better accuracy in comparison to ordinary Monte Carlo and a better visual perception in comparison to Markov chain Monte Carlo techniques with the same rendering time.

Bibtex:

@inproceedings{frolov18mixing,
    author = {Frolov, V. A. and Galaktionov, V. A. and Sangarov, V. V.},
    booktitle = {Proceedings of the 28th International Conference on Computer Graphics and Vision GraphiCon'2018},
    language = {english},
    pages = {26--31},
    publisher = {Tomsk Polytechnic University Tomsk, Russia},
    title = {Mixing ordinary and Markov chain Monte Carlo rendering techiques},
    year = {2018}
}

Abstract:

Integration layer between digital content creation software (DCCS) and rendering software in a form of specialized database is proposed in this paper. In our approach, we focus on providing fast 3D-scene updates, ability to work with large digital assets (not fitting into memory), importing and exporting arbitrary parameters, serialization, convenient debugging tools and distributed rendering. Such database can be used as means to integrate different rendering engines with DCCS and also to transfer data between different DCCS.

Bibtex:

@article{frolov18modern,
    address = {Road Town, United Kingdom},
    author = {Frolov, V. A. and Sanzharov, V. V.},
    doi = {10.1134/S0361768818040060},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {4},
    pages = {233--239},
    publisher = {Road Town, United Kingdom},
    title = {Modern Problems of Software Integration in Computer Graphics Applications and Ways to Solve Them},
    volume = {44},
    year = {2018}
}

Bibtex:

@inproceedings{sangarov18precomputed,
    author = {Sangarov, V. V. and Frolov, V. A.},
    booktitle = {Proceedings of the 28th International Conference on Computer Graphics and Vision GraphiCon'2018},
    language = {english},
    pages = {31--36},
    publisher = {Tomsk Polytechnic University Tomsk, Russia},
    title = {Precomputed Procedural content},
    year = {2018}
}

Abstract:

В данной работе впервые проведено исследование GPU-реализации алгоритма Multiplexed Metropolis Light Transport (MMLT) на репрезентативном для прикладных задач фотореалистичного рендеринга наборе 3D-сцен. Выявлены неизвестные ранее особенности и проблемы алгоритма как метода интегрирования освещённости, проявляющиеся при его реализации на массивно-параллельных процессорах и не проявляющиеся при обычной реализации на многоядерных центральных процессорах. Предложены способы решения некоторых из этих проблем. Представленное в статье сравнение с зарубежным GPU аналогом Octane показывает преимущества разработанной системы на архитектурных сценах для трудновычислимых феноменов освещённости.

Bibtex:

@article{frolov18issledovanie,
    author = {Фролов, В. А.},
    doi = {10.20948/prepr-2018-267},
    journal = {Препринт ИПМ},
    language = {russian},
    number = {267},
    pages = {1--47},
    title = {Исследование алгоритма Multiplexed Metropolis Light Transport на графических процессорах},
    year = {2018}
}

Abstract:

Предложен метод преобразования матрицы форм-факторов, позволяющий ускорять расчёт вторичного освещения методом излучательности. Рассмотрена адаптация этого метода для графических процессоров (graphics processing unit, GPU). В частности, предложено использовать DXT-текстуры для хранения матрицы форм-факторов и переупорядочивать столбцы и строки матрицы для уменьшения потерь при сжатии. Предложенные оптимизации повышают скорость работы алгоритма излучательности до 10 раз и уменьшают до 3 раз объём занимаемой памяти GPU.

Bibtex:

@article{scherbakov18matrichnye,
    address = {М.},
    author = {ЩЕРБАКОВ, А. С. and ФРОЛОВ, В. А.},
    issn = {0039-7067},
    journal = {Светотехника},
    language = {russian},
    number = {3},
    pages = {43--47},
    publisher = {М.},
    title = {Матричные преобразования для эффективной реализации алгоритма излучательности на графических процессорах},
    year = {2018}
}

Bibtex:

@inproceedings{scherbakov18obobschenie,
    author = {Щербаков, А. С. and Фролов, В. А.},
    booktitle = {Новые информационные технологии в автоматизированных системах},
    language = {russian},
    pages = {57--62},
    publisher = {ИПМ им. М.В.Келдыша},
    series = {Материалы  двадцатого  первого научно-практического  семинара},
    title = {Обобщение метода излучательности на Двунаправленные Функции Отражательной способности глянцевых материалов},
    year = {2018}
}

Bibtex:

@inproceedings{pavlov18realizatsii,
    author = {Павлов, Д. С. and Фролов, В. А.},
    booktitle = {Новые информационные технологии в автоматизированных системах},
    language = {russian},
    pages = {111--119},
    publisher = {ИПМ им. М.В.Келдыша},
    series = {Материалы  двадцатого  первого научно-практического  семинара},
    title = {Опыт реализации Gradient Domain Metropolis Light Transport},
    year = {2018}
}

Bibtex:

@inproceedings{frolov18unifitsirovannyj,
    author = {Фролов, В. А. and Галактионов, В. А.},
    booktitle = {Новые информационные технологии в автоматизированных системах},
    language = {russian},
    pages = {63--69},
    publisher = {ИПМ им. М.В.Келдыша},
    series = {Материалы  двадцатого  первого научно-практического  семинара},
    title = {Унифицированный механизм сэмплирования изображений для современных методов интегрирования освещённости на GPU},
    year = {2018}
}

Bibtex:

@inproceedings{gorokhova18combination,
    author = {Gorokhova, Svetlana and Senyukova, Olga and Bulakh, Roman and Slabouzova, Anna and Atkov, Oleg},
    booktitle = {14th International Dead Sea Symposium (IDSS) on Innovations in Cardiac Arrhythmias and Heart Failure},
    language = {english},
    pages = {74},
    title = {A Combination of ECG Intervals and Logistic Regression in Automatic Detection of Coronary Heart Disease},
    year = {2018}
}

Bibtex:

@article{gavrishchaka18advantages,
    address = {USA},
    author = {Gavrishchaka, Valeriy and Yang, Zhenyi and Miao, Rebecca and Senyukova, Olga},
    doi = {10.18178/ijmlc.2018.8.6.744},
    issn = {2010-3700},
    journal = {International Journal of Machine Learning and Computing},
    language = {english},
    number = {6},
    pages = {549--558},
    publisher = {USA},
    title = {Advantages of Hybrid Deep Learning Frameworks in Applications with Limited Data},
    volume = {8},
    year = {2018}
}

Abstract:

Segmentation of anatomical structures in cardiac MR images is an important problem because it is necessary for evaluation of morphology of these structures for diagnostic purposes. Automatic segmentation algorithm with near-human accuracy would be extremely helpful for a medical specialist. In this paper we consider such structures as endocardium and epicardium of right ventricle. We compare the performance of the best existing neural networks such as U-Net and GridNet, and propose our own modification of U-Net which implies replacement of every second convolution layer with dilated (atrous) convolution layer. Evaluation on benchmark dataset RVSC demonstrated that the proposed algorithm allows to improve the segmentation accuracy up to 6\% both for endocardium and epicardium compared to original U-Net. The algorithm also overperforms GridNet for both segmentation problems.

Bibtex:

@inproceedings{borodin18right,
    author = {Borodin, Gregory and Senyukova, Olga},
    booktitle = {Artificial Neural Networks and Machine Learning – ICANN 2018},
    doi = {10.1007/978-3-030-01421-6\_18},
    language = {english},
    pages = {179--185},
    publisher = {Springer International Publishing},
    series = {Lecture Notes in Computer Science},
    title = {Right Ventricle Segmentation in Cardiac MR Images Using U-Net with Partly Dilated Convolution},
    volume = {11140},
    year = {2018}
}

Bibtex:

@inproceedings{erofeev17preobrazovanie,
    author = {Ерофеев, М. В.},
    booktitle = {Ломоносовские чтения: Научная конференция, Москва, факультет ВМК МГУ имени М.В.Ломоносова, 17-26 апреля 2017 г. Тезисы докладов},
    language = {russian},
    pages = {114--115},
    publisher = {МАКС Пресс Москва},
    title = {Преобразование видеопоследовательностей, содержащих объекты с полупрозрачными границами, в стереоскопический формат},
    year = {2017}
}

Bibtex:

@inproceedings{kukleva17abandoned,
    author = {Kukleva, A. and Konushin, V. and Konushin, A.},
    booktitle = {Proceedings of the 27th International Conference on Computer Graphics and Vision GraphiCon'2017},
    language = {english},
    pages = {154--157},
    title = {Abandoned object detection in video sequences},
    year = {2017}
}

Bibtex:

@inproceedings{valuiskaia17automatic,
    author = {Valuiskaia, Y. and Shalnov, E. and Konushin, A. S.},
    booktitle = {Proceedings of the 27th International Conference on Computer Graphics and Vision GraphiCon'2017},
    language = {english},
    pages = {175--179},
    title = {Automatic calibration of surveillance video camera},
    year = {2017}
}

Bibtex:

@inproceedings{morozov17background,
    author = {Morozov, F. and Konushin, A.},
    booktitle = {Proceedings of the 27th International Conference on Computer Graphics and Vision GraphiCon'2017},
    language = {english},
    pages = {158--161},
    title = {Background subtraction with convolutional neural network and oversegmentation},
    year = {2017}
}

Abstract:

A new algorithm for aligning text with speech audio signals having lengths of up to several hours is proposed. The algorithm allows its quality to be effectively evaluated. The requirements on the acoustic model are not very demanding. The algorithm can be used to design an audio–video course for learning the Russian language.

Bibtex:

@article{karpukhin17constructing,
    address = {United States},
    author = {Karpukhin, I. A. and Konushin, A. S.},
    doi = {10.3103/S0278641917020030},
    issn = {0278-6419; 1934-8428},
    journal = {Moscow University Computational Mathematics and Cybernetics},
    language = {english},
    number = {2},
    pages = {97--103},
    publisher = {United States},
    title = {Constructing a speech audio–video corpus by aligning long segments of speech and text},
    volume = {41},
    year = {2017}
}

Abstract:

Known scene geometry and camera calibration parameters give important information to video content analysis systems. In this paper, we propose a novel method for camera pose estimation based on people observation in the input video captured by static camera. As opposed to previous techniques, our method can deal with false positive detections and inaccurate localization results. Specifically, the proposed method does not make any assumption about the utilized object detector and takes it as a parameter. Moreover, we do not require a huge labeled dataset of real data and train on the synthetic data only. We apply the proposed technique for camera pose estimation based on head observations. Our experiments show that the algorithm trained on the synthetic dataset generalizes to real data and is robust to false positive detections.

Bibtex:

@article{shalnov17convolutional,
    address = {Netherlands},
    author = {Shalnov, E. V. and Konushin, A. S. and Konushin, A. S.},
    doi = {10.5194/isprs-archives-XLII-2-W4-1-2017},
    issn = {0924-2716},
    journal = {ISPRS Journal of Photogrammetry and Remote Sensing},
    language = {english},
    number = {2-W4},
    pages = {1--6},
    publisher = {Netherlands},
    title = {Convolutional neural network for camera pose estimation from object detections},
    volume = {42},
    year = {2017}
}

Abstract:

Known scene geometry and camera calibration parameters give important information to video content analysis systems. In this paper, we propose a novel method for camera pose estimation based on people observation in the input video captured by static camera. As opposed to previous techniques, our method can deal with false positive detections and inaccurate localization results. Specifically, the proposed method does not make any assumption about the utilized object detector and takes it as a parameter. Moreover, we do not require a huge labeled dataset of real data and train on the synthetic data only. We apply the proposed technique for camera pose estimation based on head observations. Our experiments show that the algorithm trained on the synthetic dataset generalizes to real data and is robust to false positive detections.

Bibtex:

@article{shalnov17convolutionalneural,
    address = {Netherlands},
    author = {Shalnov, E. V. and Konushin, A. S. and Konushin, A. S.},
    doi = {10.5194/isprs-archives-XLII-2-W4-1-2017},
    issn = {0924-2716},
    journal = {ISPRS Journal of Photogrammetry and Remote Sensing},
    language = {english},
    number = {2-W4},
    pages = {1--6},
    publisher = {Netherlands},
    title = {Convolutional neural network for camera pose estimation from object detections},
    volume = {42},
    year = {2017}
}

Abstract:

In this work we investigate the problem of people recognition by their gait. For this task, we implement deep learning approach using the optical flow as the main source of motion information and combine neural feature extraction with the additional embedding of descriptors for representation improvement. In order to find the best heuristics, we compare several deep neural network architectures, learning and classification strategies. The experiments were made on two popular datasets for gait recognition, so we investigate their advantages and disadvantages and the transferability of considered methods.

Bibtex:

@article{sokolova17recognition,
    address = {Netherlands},
    author = {Sokolova, A. and Konushin, A.},
    doi = {10.5194/isprs-archives-XLII-2-W4-207-2017},
    issn = {0924-2716},
    journal = {ISPRS Journal of Photogrammetry and Remote Sensing},
    language = {english},
    number = {2-W4},
    pages = {207--212},
    publisher = {Netherlands},
    title = {Gait recognition based on convolutional neural networks},
    volume = {42},
    year = {2017}
}

Bibtex:

@inproceedings{konushin17recognition,
    author = {Konushin, A. and Arseev, A.},
    booktitle = {Proceedings of the 27th International Conference on Computer Graphics and Vision GraphiCon'2017},
    language = {english},
    pages = {151--153},
    title = {Gait recognition based on relative distance and motion features},
    year = {2017}
}

Abstract:

The paper considers a problem of multiple person tracking. We present the algorithm to automatic people tracking on surveillance videos recorded by static cameras. Proposed algorithm is an extension of approach based on tracking-by-detection of people heads and data association using Markov chain Monte Carlo (MCMC). Short track fragments (tracklets) are built by local tracking of people heads. Tracklet postprocessing and accurate results interpolation were shown to reduce number of false positives. We use position deviations of tracklets and revised entry/exit points factor to separate pedestrians from false positives. The paper presents a new method to estimate body position, that increases precision of tracker. Finally, we switched HOG-based detector to cascade one. Our evaluation shows proposed modifications significantly increase tracking quality.

Bibtex:

@article{kuplyakov17markov,
    address = {Road Town, United Kingdom},
    author = {Kuplyakov, D. and Shalnov, E. and Konushin, A.},
    doi = {10.1134/S0361768817040053},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {4},
    pages = {224--229},
    publisher = {Road Town, United Kingdom},
    title = {Markov chain Monte Carlo based video tracking algorithm},
    volume = {43},
    year = {2017}
}

Abstract:

В статье представлен алгоритм автоматического сопровождения людей в видеопоследовательности, полученной со статичной камеры видеонаблюдения. Предложенный алгоритм является развитием подхода на основе объединения коротких фрагментов траекторий (треклетов) с помощью метода Монте-Карло для Марковской цепи (MCMC). Короткие фрагменты траекторий строятся с помощью визуального сопровождения голов людей, выделенных детектором. Обработка треклетов и точная интерполяция результатов позволили уменьшить число ложных обнаружений. Алгоритм использует среднее отклонение оценок положения и новый подход к учету границы входа/выхода, чтобы отделить обнаружения людей от ложных обнаружений. Представлен новый метод оценки положения тела, который позволил улучшить точность сопровождения. Также были проведены эксперименты с заменой детектора на основе гистограмм ориентированных градиентов на каскадный детектор. Экспериментальное сравнение показало, что представленные модификации значительно улучшили качество сопровождения.

Bibtex:

@article{kupljakov17algoritm,
    author = {Купляков, Д. А. and Шальнов, Е. В. and Конушин, А. С.},
    issn = {0132-3474},
    journal = {Программирование},
    language = {russian},
    number = {4},
    pages = {13--21},
    title = {Алгоритм сопровождения людей в видео на основе метода Монте-Карло для марковских цепей},
    year = {2017}
}

Bibtex:

@inproceedings{ljutov17glubinnye,
    author = {Лютов, В. and Конушин, А.},
    booktitle = {Proceedings of the 27th International Conference on Computer Graphics and Vision GraphiCon'2017},
    language = {russian},
    pages = {217--222},
    title = {Глубинные двоичные дескрипторы изображения человека для его повторной идентификации и сопровождения в видео},
    year = {2017}
}

Bibtex:

@article{shalnov17ispolzovanie,
    author = {Шальнов, Е. В. and Конушин, А. С.},
    doi = {10.15827/0236-235X.117.106-111},
    issn = {0236-235X},
    journal = {Программные продукты и системы},
    language = {russian},
    number = {1},
    pages = {106--111},
    title = {Использование геометрии сцены для увеличения точности детекторов},
    volume = {30},
    year = {2017}
}

Abstract:

Данная работа посвящена задаче распознавания людей по лицу в видеопоследовательности. В работе предложена нейросетевая модель, которая для входного набора изображений лица человека строит компактное признаковое представление фиксированной размерности. Предложенная модель состоит из двух частей: модуль распознавания по изображению лица и модуль оценки качества изображения лица. Признаковые представления кадров из входного набора, полученные в результате работы модуля распознавания, агрегируются с учетом их полезности, которая оценивается модулем оценки качества. Визуальный анализ выявил, что предложенная нейронная сеть учится использовать больше полезной информации с изображений высокого качества и меньше – с размытых или перекрытых изображений. Экспериментальная оценка на базах YouTube Faces и IJB-A показала, что предложенный метод объединения признаков на основе оценок полезности изображений позволяет повысить качество распознавания по сравнению с базовыми методами агрегации.

Bibtex:

@article{nikitin17nejrosetevaja,
    author = {Никитин, М. Ю. and Конушин, В. С. and Конушин, А. С.},
    doi = {10.18287/2412-6179-2017-41-5-732-742},
    issn = {0134-2452},
    journal = {Компьютерная оптика},
    language = {russian},
    number = {5},
    pages = {732--742},
    title = {Нейросетевая модель распознавания человека по лицу в видеопоследовательности с оценкой полезности кадров},
    volume = {41},
    year = {2017}
}

Abstract:

This work aims to apply visual-attention modeling to attention-based video compression. During our comparison we found that eye-tracking data collected even from a single observer outperforms existing automatic models by a significant margin. Therefore, we offer a semiautomatic approach: using computer-vision algorithms and good initial estimation of eye-tracking data from just one observer to produce high-quality saliency maps that are similar to multi-observer eye tracking and that are appropriate for practical applications. We propose a simple algorithm that is based on temporal coherence of the visual-attention distribution and requires eye tracking of just one observer. The results are as good as an average gaze map for two observers. While preparing the saliency-model comparison, we paid special attention to the quality-measurement procedure. We observe that many modern visual-attention models can be improved by applying simple transforms such as brightness adjustment and blending with the center-prior model. The novel quality-evaluation procedure that we propose is invariant to such transforms. To show the practical use of our semiautomatic approach, we developed a saliency-aware modification of the x264 video encoder and performed subjective and objective evaluations. The modified encoder can serve with any attention model and is publicly available.

Bibtex:

@inproceedings{lyudvichenko17semiautomatic,
    author = {Lyudvichenko, V. and Erofeev, M. and Gitman, Y. and Vatolin, D.},
    booktitle = {2017 IEEE 13th International Conference on Intelligent Computer Communication and Processing},
    doi = {10.1109/ICCP.2017.8117038},
    language = {english},
    pages = {403--410},
    title = {A semiautomatic saliency model and its application to video compression},
    year = {2017}
}

Abstract:

Objective technical quality assessment of stereoscopic movies is crucial for timely detection and correction of potentially problematic scenes in order to decrease the fraction of viewers who might experience visual discomfort. Despite the prior advancements in the field, some recently released movies still contain scenes with significant technical issues. For instance, the movie Mermaid, which grossed over \$500 million in box office, contains both artifacts common for natively captured S3D content and artifacts specific to 2D-to-3D conversion. In addition to previously described technical quality metrics this paper introduces our improved approach to measuring sharpness mismatch between stereoscopic views, as well as a detailed analysis of 10 recently released Chinese S3D movies using 7 different quality metrics. The analyzed 10 movies are compared with general stereoscopic movie quality trends evaluated on over hundred S3D movies, produced mostly in Europe and USA.

Bibtex:

@inproceedings{vatolin17sharpness,
    address = {United States},
    author = {Vatolin, Dmitriy and Bokov, Alexander},
    booktitle = {Proceedings of Stereoscopic Displays and Applications XXVIII},
    doi = {10.2352/ISSN.2470-1173.2017.5.SDA-340},
    language = {english},
    pages = {137--144},
    publisher = {United States},
    title = {Sharpness Mismatch and 6 Other Stereoscopic Artifacts Measured on 10 Chinese S3D Movies},
    year = {2017}
}

Abstract:

Today, numerous movies are produced in stereoscopic format. Despite the improvement in stereo technology, stereoscopic artifacts that cause headaches and other viewer discomfort continue to appear even in high-budget films. Existing automatic quality-control algorithms can detect distortions in stereoscopic images, but they fail to account for a viewer’s subjective perception of those distortions. We propose a method of automatic subjective quality evaluation that uses technical parameters of stereoscopic scenes. It is based on subjective scores and brain-activity measurements using electroencephalography (EEG) to assess viewer discomfort. We conducted a series of experiments with active and passive stereo cinema technology. An audience of 302 participants watched 60 video sequences from stereoscopic movies containing artificially added geometric, color and temporal artifacts. Our analysis of the data revealed the dependencies between the degree of viewer discomfort and the intensity of the distortions.Scenes with temporal mismatch between the stereoscopic views caused the most discomfort among viewers. Future work will focus on creating models based on this data and using them to predict audience discomfort caused by watching stereoscopic movies.

Bibtex:

@inproceedings{antsiferova17influence,
    author = {Antsiferova, Anastasia and Vatolin, Dmitriy},
    booktitle = {2017 International Conference on 3D Immersion (IC3D)},
    doi = {10.1109/IC3D.2017.8251897},
    language = {english},
    pages = {1--8},
    publisher = {IEEE},
    title = {The influence of 3D video artifacts on discomfort of 302 viewers},
    year = {2017}
}

Abstract:

We propose an efficient method for reconstructing background fragments that are visible in at least one frame of the input video sequence; this method can aid in filling disoccluded areas during view synthesis. We make no assumptions regarding background or camera motion except that it is smooth. The proposed method can therefore handle a wide variety of scenes captured with a free-moving camera. Our approach relies on formulating the problem as frame-by-frame global energy minimization, which we can efficiently solve given that the previous frame provides a good initial approximation. We compare our approach with a commercial background-reconstruction tool for view synthesis and with a general-purpose video-completion algorithm.

Bibtex:

@inproceedings{bokov17toward,
    author = {Bokov, Alexander and Vatolin, Dmitriy},
    booktitle = {IEEE International Conference on Multimedia \& Expo Workshops (ICMEW)},
    doi = {10.1109/ICMEW.2017.8026297},
    language = {english},
    pages = {37--42},
    publisher = {Hong Kong, China},
    title = {Toward efficient background reconstruction for 3D-view synthesis in dynamic scenes},
    year = {2017}
}

Bibtex:

@article{groholskij17avtomaticheskoe,
    address = {Москва},
    author = {Грохольский, С. Д. and Лаврушкин, С. В. and Звездаков, С. В. and Ватолин, Д. С.},
    issn = {2307-8162},
    journal = {International Journal of Open Information Technologies},
    language = {russian},
    number = {5},
    pages = {1--8},
    publisher = {Москва},
    title = {Автоматическое обнаружение и оценка цветовых искажений с использованием карт доверия},
    volume = {5},
    year = {2017}
}

Bibtex:

@inproceedings{molodetskih17algoritm,
    author = {Молодецких, И. А. and Ватолин, Д. С.},
    booktitle = {Инновационные технологии в кинематографе и образовании: IV Международная научно-практическая конференция},
    language = {russian},
    pages = {75--85},
    publisher = {ВГИК Москва},
    title = {Алгоритм автоматического построения карт перемонтажа разных версий фильма},
    year = {2017}
}

Bibtex:

@inproceedings{kondranin17algoritm,
    author = {Кондранин, Д. С. and Ватолин, Д. С.},
    booktitle = {Инновационные технологии в кинематографе и образовании: IV Международная научно-практическая конференция},
    language = {russian},
    pages = {66--74},
    publisher = {ВГИК Москва},
    title = {Алгоритм выбора областей кадров, наиболее подходящих для демонстрации дефектов сжатия},
    year = {2017}
}

Bibtex:

@article{bokov17vosstanovlenie,
    author = {Боков, А. А. and Ватолин, Д. С.},
    issn = {1684-2634},
    journal = {Цифровая обработка сигналов},
    language = {russian},
    number = {3},
    pages = {13--21},
    title = {Восстановление фона в видео с использованием непараметрической модели движения и покадрового уточнения},
    year = {2017}
}

Bibtex:

@inproceedings{vatolin17vosstanovlenie,
    author = {Ватолин, Д. С. and Боков, А. А.},
    booktitle = {Ломоносовские чтения: Научная конференция, Москва, факультет ВМК МГУ имени М.В.Ломоносова, 17-26 апреля 2017 г. Тезисы докладов},
    language = {russian},
    pages = {115--116},
    publisher = {МАКС Пресс Москва},
    title = {Восстановление фона в задаче конвертации видио в стереоскопический формат},
    year = {2017}
}

Bibtex:

@article{antsiferova17issledovanie,
    author = {Анциферова, А. В. and Ватолин, Д. С. and Звездаков, С. В.},
    issn = {1991-3400},
    journal = {Мир техники кино},
    language = {russian},
    number = {2},
    pages = {8--12},
    title = {Исследование влияния геометрических, цветовых и временных искажений стереоскопического видео на дискомфорт зрителей},
    volume = {11},
    year = {2017}
}

Bibtex:

@inproceedings{antsiferova17issledovanievlijanija,
    author = {Анциферова, А. В. and Ватолин, Д. С. and Звездаков, С. В.},
    booktitle = {Запись и воспроизведение объемных изображений в кинематографе и других областях: IX Международная научно-практическая конференция},
    language = {russian},
    pages = {88--102},
    publisher = {ВГИК Москва},
    title = {Исследование влияния геометрических, цветовых и временных искажений стереоскопического видео на дискомфорт зрителей},
    year = {2017}
}

Bibtex:

@inproceedings{zvezdakov17postroenie,
    author = {Звездаков, С. В. and Ватолин, Д. С.},
    booktitle = {Инновационные технологии в кинематографе и образовании: IV Международная научно-практическая конференция},
    language = {russian},
    pages = {56--65},
    publisher = {ВГИК Москва},
    title = {Построение модели видеокодека x264},
    year = {2017}
}

Bibtex:

@inproceedings{frolov17concept,
    author = {Frolov, V. F. and Sangarov, V. S. and Galaktionov, V. A.},
    booktitle = {труды 27 международной конференции GraphiCon-2017},
    language = {english},
    pages = {28--31},
    publisher = {Пермский государственный национальный исследовательский университет},
    title = {A concept for database oriented 3D graphics engine infrastructure},
    year = {2017}
}

Abstract:

We propose a novel approach to implement radiosity on GPU with specific optimizations via form-factor matrix transformations. The proposed transformations enable to reduce the number of computations for multiple-bounce global illumination and apply DXT compression (with subsequent hardware decompression when reading form-factors on GPU). Our implementation is 10 times faster running and requires 3 times less memory than the naive radiosity GPU implementation.

Bibtex:

@inproceedings{shcherbakov17accelerating,
    author = {Shcherbakov, Alexandr and Frolov, Vladimir},
    booktitle = {WSCG'2017 Full papers proceedings},
    language = {english},
    pages = {99--105},
    publisher = {Computer Science Research Notes 2701 Pilsen, Czech Republic},
    series = {2701},
    title = {Accelerating Radiosity on GPUs},
    year = {2017}
}

Abstract:

Solutions to the key problems of Metropolis light transport implementation on GPUs are proposed. A “burn-in” method relying on the ordinary Monte Carlo method, owing to which the “startup bias” is significantly reduced, is suggested. Memory optimizations methods (including multiple proposal Metropolis light transport) are proposed, and technical aspects of efficient Metropolis light transport implementation on GPUs are discussed.

Bibtex:

@article{frolov17memorycompact,
    address = {Road Town, United Kingdom},
    author = {Frolov, V. A. and Galaktionov, V. A.},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {3},
    pages = {196--203},
    publisher = {Road Town, United Kingdom},
    title = {Memory-Compact Metropolis Light Transport on GPUs},
    volume = {43},
    year = {2017}
}

Abstract:

В данной работе предложен метод совмещения обыкновенного метода Монте-Карло на основе двунаправленной трассировки путей (BPT) и Монте-Карло по схеме марковских цепей на основе метода Multiplexed Metropolis Light Transport (MMLT). Совмещение осуществляется в пространстве изображения при помощи попиксельной маски, полученной на основе анализа шума в изображении, сгенерированном двунаправленной трассировкой путей. По итогам совмещения получено 2 основных результата — (1) увеличение точности расчёта по сравнению с BPT при одинаковом времени расчёта и (2) визуальное улучшение изображения в предложенном методе по сравнению с MMLT при одинаковом времени расчёта.

Bibtex:

@article{frolov17izbiratelnoe,
    address = {Москва},
    author = {Фролов, В. А.},
    doi = {10.20948/prepr-2017-116},
    issn = {2071-2898; 2071-2901},
    journal = {Препринты ИПМ им. М.В. Келдыша},
    language = {russian},
    number = {116},
    pages = {1--34},
    publisher = {Москва},
    title = {Избирательное применение Metropolis Light Transport для трудновычислимых феноменов освещённости},
    year = {2017}
}

Bibtex:

@inproceedings{scherbakov17uskorenie,
    author = {Щербаков, А. С. and Фролов, В. А.},
    booktitle = {труды 27 международной конференции GraphiCon-2017},
    language = {russian},
    pages = {63--67},
    publisher = {Пермский государственный национальный исследовательский университет},
    title = {Ускорение алгоритма излучательности на графических процессорах},
    year = {2017}
}

Bibtex:

@inproceedings{senyukova17automated,
    author = {Senyukova, O. and Brotikovskaya, D. and Gorokhova, S. and Tebenkova, E.},
    booktitle = {Proceedings of the 9th International Joint Conference on Computational Intelligence, Funchal, Madeira, Portugal},
    doi = {10.5220/0006518203600366},
    language = {english},
    pages = {360--366},
    title = {Automated Diagnostic Model Based on Heart Tissue Isoline Map Analysis},
    year = {2017}
}

Bibtex:

@inproceedings{senyukova17multiexpert,
    author = {Senyukova, O. and Gavrishchaka, V. and Tulnova, K.},
    booktitle = {2017 Evolving and Adaptive Intelligent Systems (EAIS), Ljubljana, Slovenia},
    doi = {10.1109/EAIS.2017.7954824},
    language = {english},
    pages = {1--8},
    title = {Multi-expert evolving system for objective psychophysiological monitoring and fast discovery of effective personalized therapies},
    year = {2017}
}

Bibtex:

@article{senjukova17avtomatizirovannaja,
    author = {Сенюкова, О. В. and Гаврищака, В. В. and Тульнова, К. С. and Монин, А. Г.},
    journal = {Медицинская психология в России: электрон. науч. журн},
    language = {russian},
    number = {6(47)},
    title = {Автоматизированная система для психофизиологического мониторинга и подбора персонализированной терапии},
    volume = {9},
    year = {2017}
}

Bibtex:

@inproceedings{morozov16background,
    author = {Morozov, F. and Konushin, A. S.},
    booktitle = {Proceedings of the 26th International Conference on Computer Graphics and Vision GraphiCon'2016},
    language = {english},
    pages = {445--447},
    series = {GraphiCon},
    title = {Background subtraction using a convolutional neural network},
    year = {2016}
}

Abstract:

A method is proposed for estimating the position of people in a scene when their head locations are known in the image plane. An extension of the approach is presented for processing several observations of the same person. It is shown that the algorithm proposed can be incorporated in the existing tracking methods involving a video from a static camera.

Bibtex:

@article{shal’nov16estimation,
    address = {Road Town, United Kingdom},
    author = {Shal’nov, E. V. and Gringauz, A. D. and Konushin, A. S.},
    doi = {10.1134/S0361768816060062},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {6},
    pages = {361--366},
    publisher = {Road Town, United Kingdom},
    title = {Estimation of the people position in the world coordinate system for video surveillance},
    volume = {42},
    year = {2016}
}

Bibtex:

@inproceedings{arseev16smoke,
    author = {Arseev, Sergey and Konushin, Anton and Konushin, Vadim and Lukina, Tatiana and Sindeev, Mikhail},
    booktitle = {Proceedings of the 26th International Conference on Computer Graphics and Vision GraphiCon'2016},
    language = {english},
    pages = {527--530},
    series = {GraphiCon},
    title = {Fire and Smoke Detection in Video Sequences},
    year = {2016}
}

Bibtex:

@inproceedings{kuplyakov16further,
    author = {Kuplyakov, Denis and Shalnov, Evgeny and Konushin, Anton},
    booktitle = {Proceedings of the 26th International Conference on Computer Graphics and Vision GraphiCon'2016},
    language = {english},
    pages = {440--444},
    series = {GraphiCon},
    title = {Further Improvement on an MCMC-based Video Tracking Algorithm},
    year = {2016}
}

Abstract:

Данная работа посвящена задаче фильтрации ложноположительных обнаружений лиц людей в видеопотоках систем видеонаблюдения. Предлагается два подхода для подавления ложноположительных обнаружений в фоновых участках кадра: первый заключается в адаптации детектора под наблюдаемый видеопоток, а второй представляет собой постобработку выхода детектора за счёт анализа частоты обнаружения похожих частей кадра. Мы используем в качестве базового каскадный детектор, но метод может быть применён к другим алгоритмам. Экспериментальное оценивание показывает, что предложенные методы улучшают и точность, и полноту, при этом время работы детектора сокращается на 47 \%.

Bibtex:

@article{sergeev16podavlenie,
    author = {Сергеев, А. Е. and Конушин, А. С. and Конушин, В. С.},
    doi = {10.18287/2412-6179-2016-40-6-958-967},
    issn = {0134-2452},
    journal = {Компьютерная оптика},
    language = {russian},
    number = {6},
    pages = {958--967},
    title = {Подавление ложноположительных обнаружений лиц в видеопотоках систем видеонаблюдения},
    volume = {40},
    year = {2016}
}

Abstract:

Представлена новая открытая база изображений дорожных знаков. База предназначена для обучения и тестирования алгоритмов распознавания дорожных знаков. Описывается структура и правила работы с базой, проводится сравнение с другими базами дорожных знаков. На базе проведена экспериментальная оценка современных алгоритмов выделения и классификации дорожных знаков, которая показала, что существующие алгоритмы распознавания большого класса знаков не достигают требуемой для ряда практических приложений полноты и точности.

Projects:

Bibtex:

@article{shahuro16rossijskaja,
    author = {Шахуро, В. И. and Конушин, А. С.},
    doi = {10.18287/2412-6179-2016-40-2-294-300},
    issn = {0134-2452},
    journal = {Компьютерная оптика},
    language = {russian},
    number = {2},
    pages = {294--300},
    title = {Российская база изображений автодорожных знаков},
    volume = {40},
    year = {2016}
}

Abstract:

The degree of visual discomfort caused by watching stereoscopic scenes with channel mismatch is investigated and predicted. A scene with channel order mismatch is one in which the right and left views are swapped. A way of finding channelmismatch is used to analyze 105 3D films; the scenes found in this analysis are used for experimental study of the visual discomfort caused by channelmismatch. The experimental results are used to construct samplingwith a reference pattern. This sampling is used to learn different regression analysis algorithms, and the bestway of predicting visual discomfort caused by channel mismatch is chosen.

Bibtex:

@article{vatolin16investigating,
    address = {United States},
    author = {Vatolin, D. S. and Lavrushkin, S. V.},
    doi = {10.3103/s0278641916040075},
    issn = {0278-6419; 1934-8428},
    journal = {Moscow University Computational Mathematics and Cybernetics},
    language = {english},
    number = {4},
    pages = {185--191},
    publisher = {United States},
    title = {Investigating and predicting the perceptibility protect of channel mismatch in stereoscopic video},
    volume = {40},
    year = {2016}
}

Abstract:

Owing to the movie industry's wide use of 2D-to-3D conversion techniques, the problem of synthesizing stereoscopic views using depth-image-based rendering (DIBR) is extremely important. A major challenge for DIBR is processing semitransparent edges near depth-map discontinuities. Existing approaches either can only deal with simple cases in which the background behind the object does not change significantly, or they segment the input image into two layers. Unfortunately, such segmentation is impossible to carry out correctly for depth discontinuities formed by multiple objects, leading to visual artifacts in the generated views. Our proposed method of multilayer semitransparent-edge processing avoids these limitations and outperforms competitors in a subjective visual-quality comparison of the synthesized views.

Bibtex:

@inproceedings{erofeev16multilayer,
    author = {Erofeev, M. and Vatolin, D.},
    booktitle = {2016 International Conference on 3D Imaging (IC3D)},
    doi = {10.1109/IC3D.2016.7823459},
    language = {english},
    pages = {1--7},
    publisher = {Liege},
    title = {Multilayer semitransparent-edge processing for depth-image-based rendering},
    year = {2016}
}

Abstract:

Channel mismatch — i.e., swapped left and right views — can cause major visual discomfort when viewing stereoscopic content, but detecting it automatically remains a challenge. We present a novel method for detecting channel mismatch that significantly outperforms a prior approach when analyzing a dataset of 1,000 video clips. Moreover, this method enabled detection of 65 scenes exhibiting channel mismatch in 105 real S3D movies. Perceived discomfort can vary greatly between different scenes with channel mismatch depending on a number of factors. We conducted a study involving 59 participants to compute subjective discomfort scores for 56 scenes with channel mismatch. In this paper, we propose a model to predict these scores on the basis of scene characteristics.

Bibtex:

@inproceedings{bokov16toward,
    author = {Bokov, Alexander and Lavrushkin, Sergey and Erofeev, Mikhail and Vatolin, Dmitriy and Fedorov, Alexey},
    booktitle = {2016 International Conference on 3D Imaging (IC3D)},
    doi = {10.1109/IC3D.2016.7823462},
    language = {english},
    pages = {1--7},
    publisher = {Liege},
    title = {Toward fully automatic channel-mismatch detection and discomfort prediction for S3D video},
    year = {2016}
}

Abstract:

In this paper we present a large-scale analysis of S3D-movie technical quality spanning a large portion of stereoscopic-cinema history. We evaluated 105 Blu-ray 3D releases, including titles like the 1954 classic Dial M for Murder, as well as contemporary stereoscopic productions like Life of Pi and The Great Gatsby. The analysis is based on objective quality metrics designed to detect specific types of artifacts, including swapped channels, inconsistencies between the stereoscopic views (color, sharpness and geometric as well as temporal asynchrony) and many others. The main challenges we had to overcome were the enormous amount of computational resources and disk space that such analyses require as well as algorithmic difficulties in developing some of the more advanced objective quality metrics. Our study clarifies the quality trends and problems of S3D movie production in general and provides a better understanding of how effectively quality control has been applied to particular movies compared with the overall trend.

Bibtex:

@inproceedings{vatolin16trends,
    author = {Vatolin, Dmitriy and A, Bokov Alexander and Erofeev, Mikhail and Napadovsky, Vyacheslav},
    booktitle = {Proceedings of Stereoscopic Displays and Applications XXVII},
    doi = {10.2352/ISSN.2470-1173.2016.5.SDA-439},
    language = {english},
    pages = {SDA-439.1--SDA-439.10},
    publisher = {San Francisco},
    title = {Trends in S3D-Movie Quality Evaluated on 105 Films Using 10 Metrics},
    year = {2016}
}

Bibtex:

@inproceedings{zvezdakov16avtomaticheskaja,
    author = {Звездаков, С. В. and Боков, А. А. and Ватолин, Д. С.},
    booktitle = {Запись и воспроизведение объемных изображений в кинематографе и других областях: VIII Международная научно-практическая конференция},
    language = {russian},
    pages = {198--211},
    title = {Автоматическая многоклассовая классификация планов стереовидео по методу производства},
    year = {2016}
}

Bibtex:

@inproceedings{antsiferova16avtomaticheskij,
    author = {Анциферова, А. В. and Ватолин, Д. С.},
    booktitle = {Запись и воспроизведение объемных изображений в кинематографе и других областях: VIII Международная научно-практическая конференция},
    language = {russian},
    pages = {106--122},
    title = {Автоматический метод оценки степени усталости от просмотра 3D-видео},
    year = {2016}
}

Bibtex:

@article{vatolin16issledovanie,
    address = {М.},
    author = {Ватолин, Д. С. and Лаврушкин, С. В.},
    issn = {0137-0782},
    journal = {Вестник Московского университета. Серия 15: Вычислительная математика и кибернетика},
    language = {russian},
    pages = {40--46},
    publisher = {М.},
    title = {Исследование и предсказание заметности перепутанных ракурсов в стереовидео},
    volume = {15},
    year = {2016}
}

Bibtex:

@article{vatolin16kratkij,
    author = {Ватолин, Д. С.},
    issn = {1991-3400},
    journal = {Мир техники кино},
    language = {russian},
    number = {1},
    pages = {5--7},
    title = {Краткий обзор современных практических подходов к работе с четырёхмерным видео},
    year = {2016}
}

Abstract:

Задача матирования состоит в разделении данного изображения или видео на передний план, задний план и карту прозрачности переднего плана. Построение такого разделения важно для решения многих задач редактирования и обработки видео, таких как: замена фона, применение спецэффекта только к фону либо только к переднему плану, построение стереоизображения. Предлагается метод матирования видеопоследовательностей, основанный на методе матирования изображений Learning Based Matting. Описывается способ модификации базового метода, позволивший использовать информацию о фрагментах фона, скрытых объектом переднего плана, в качестве дополнительных данных; а также способ итеративной фильтрации последовательности карт прозрачности. Приводятся результаты сравнения предложенного алгоритма с 11 аналогами, иллюстрирующие преимущества предложенного подхода.

Bibtex:

@article{erofeev16matirovanie,
    author = {Ерофеев, М. В. and Ватолин, Д. С.},
    issn = {1684-2634},
    journal = {Цифровая обработка сигналов},
    language = {russian},
    number = {3},
    pages = {19--25},
    title = {Матирование видеопоследовательностей с использованием восстановленного фона},
    year = {2016}
}

Bibtex:

@article{bokov16metodika,
    author = {Боков, А. А. and Ватолин, Д. С.},
    issn = {1684-2634},
    journal = {Цифровая обработка сигналов},
    language = {russian},
    number = {3},
    pages = {26--33},
    title = {Методика объективной оценки качества восстановления фона в видео},
    year = {2016}
}

Abstract:

Задача построения стереоскопических ракурсов на основе исходного моноскопического изображения и карты глубины крайне актуальна благодаря широкому применению техник преобразования моноскопических видеопоследовательностей в стереоскопический формат при производстве современных фильмов. Существенной проблемой, возникающей при ее решении, является обработка пикселов, в которых смешаны цвета нескольких объектов, расположенных на разном удалении от зрителя. Существующие подходы к ее решению либо не способны обработать существенное изменение цвета фона за объектом на новом ракурсе, либо выполняют сегментацию изображения на два слоя, что, к сожалению, не всегда возможно сделать корректно. Предлагаемый в статье метод построения новых ракурсов лишен вышеупомянутых недостатков благодаря использованию многослойного представления карты глубины. Как показывает сравнение с существующими аналогами, предложенный метод позволяет получить результаты более высокого визуального качества.

Bibtex:

@article{erofeev16mnogoslojnoe,
    address = {Москва},
    author = {Ерофеев, Михаил Викторович and Ватолин, Дмитрий Сергеевич},
    issn = {2307-8162},
    journal = {International Journal of Open Information Technologies},
    language = {russian},
    number = {8},
    publisher = {Москва},
    title = {Многослойное решение проблемы полупрозрачных границ при построении стереоскопических изображений},
    volume = {4},
    year = {2016}
}

Bibtex:

@article{vatolin16perehod,
    author = {Ватолин, Д. С. and Боков, А. А.},
    issn = {1991-3400},
    journal = {Мир техники кино},
    language = {russian},
    number = {3},
    pages = {14--23},
    title = {Переход индустрии 3D-фильмов к конвертации и методы оценки её качества},
    year = {2016}
}

Bibtex:

@inproceedings{vatolin16perehodindustrii,
    author = {Ватолин, Д. С. and Боков, А. А.},
    booktitle = {Запись и воспроизведение объемных изображений в кинематографе и других областях: VIII Международная научно-практическая конференция},
    language = {russian},
    pages = {37--55},
    title = {Переход индустрии стереофильмов к конвертации и методы оценки её качества},
    year = {2016}
}

Bibtex:

@article{vatolin16rendering,
    author = {Ватолин, Д. С.},
    issn = {1991-3400},
    journal = {Мир техники кино},
    language = {russian},
    number = {4},
    pages = {14--14},
    title = {Рендеринг фокуса, разрешения, формы диафрагмы, формы затвора пленоптических камер},
    volume = {10},
    year = {2016}
}

Abstract:

published in Programmirovanie

Bibtex:

@article{frolov16overhead,
    address = {Road Town, United Kingdom},
    author = {Frolov, V. A. and Galaktionov, V. A.},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {6},
    pages = {382--387},
    publisher = {Road Town, United Kingdom},
    title = {Low Overhead Path Regeneration},
    volume = {42},
    year = {2016}
}

Bibtex:

@inproceedings{scherbakov16avtomaticheskoe,
    author = {Щербаков, А. and Фролов, В.},
    booktitle = {Сборник трудов Графикон 2016},
    language = {russian},
    pages = {34--38},
    publisher = {ННГАСУ},
    title = {Автоматическое упрощение геометрии для расчёта вторич- ной освещенности методом излучательности},
    year = {2016}
}

Bibtex:

@inproceedings{nikerova16gibridnyj,
    author = {Никерова, О. and Фролов, В. and Волобой, А.},
    booktitle = {Сборник трудов Графикон 2016},
    language = {russian},
    publisher = {ННГАСУ},
    title = {Гибридный подход к фильтрации результатов Монте-Карло трассировки лучей по нескольким изображениям},
    year = {2016}
}

Abstract:

В данной работе предлагается новый подход к распределенному рендерингу как на GPU, так и на CPU, использующий идеи grid вычислений. Предложена новая модель поведения вычислительных узлов на основе акторного подхода. Проведено численное исследование свойств модели.

Bibtex:

@article{sanzharov16issledovanie,
    author = {Санжаров, В. В. and Фролов, В. А.},
    journal = {Препринт ИПМ},
    language = {russian},
    number = {114},
    pages = {1--22},
    title = {Исследование масштабируемости распределённых рендер-систем на основе алгоритмов адаптивной трассировки путей и Metropolis Light Transport в гетерогенных сетях},
    year = {2016}
}

Bibtex:

@inproceedings{frolov16issledovanie,
    author = {Фролов, В. and Галактионов, В.},
    booktitle = {Сборник трудов Графикон 2016},
    language = {russian},
    pages = {227--231},
    publisher = {ННГАСУ},
    title = {Исследование особенностей Metropolis Light Transport на GPU},
    year = {2016}
}

Bibtex:

@inproceedings{sanzharov16modelirovanie,
    author = {Санжаров, В. В. and Фролов, В. А.},
    booktitle = {Сборник трудов Графикон 2016},
    language = {russian},
    pages = {248--252},
    publisher = {ННГАСУ},
    title = {Моделирование распределенных вычислений для задачи синтеза фотореалистичных изображений на основе акторов и GRID технологий},
    year = {2016}
}

Bibtex:

@article{frolov16regeneratsija,
    author = {Фролов, В. А. and Галактионов, В. А.},
    issn = {0132-3474},
    journal = {Программирование},
    language = {russian},
    number = {6},
    pages = {67--74},
    title = {Регенерация путей с низкими накладными расходами},
    year = {2016}
}

Bibtex:

@article{senyukova16anatomical,
    address = {Road Town, United Kingdom},
    author = {Senyukova, O. and Zubov, A.},
    doi = {10.1134/S0361768816060050},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {6},
    pages = {356--360},
    publisher = {Road Town, United Kingdom},
    title = {Full Anatomical Labeling of Magnetic Resonance Images of Human Brain by Registration with Multiple Atlases},
    volume = {46},
    year = {2016}
}

Bibtex:

@inproceedings{senyukova16generic,
    author = {Senyukova, O. and Gavrishchaka, V. and Sasonko, M. and Gurfinkel, Y. and Gorokhova, S. and Antsygin, N.},
    booktitle = {Computational Collective Intelligence: 8th International Conference, ICCCI 2016, Halkidiki, Greece, September 28-30, 2016. Proceedings, Part I},
    doi = {10.1007/978-3-319-45243-2\_18},
    language = {english},
    pages = {197--207},
    publisher = {Springer International Publishing Cham, Switzerland},
    series = {Lecture Notes in Computer Science},
    title = {Generic Ensemble-Based Representation of Global Cardiovascular Dynamics for Personalized Treatment Discovery and Optimization},
    volume = {9875},
    year = {2016}
}

Bibtex:

@inproceedings{senyukova16segmentation,
    author = {Senyukova, O. and Zobnin, D.},
    booktitle = {Proceedings of the 26th International Conference on Computer Graphics and Vision GraphiCon'2016},
    language = {english},
    pages = {161--165},
    publisher = {Nizhny Novgorod, Russia},
    title = {Segmentation of brain MRI with Fast and Efficient Registration Algorithm},
    year = {2016}
}

Bibtex:

@article{senjukova16polnaja,
    author = {Сенюкова, О. В. and Зубов, А. Ю.},
    issn = {0132-3474},
    journal = {Программирование},
    language = {russian},
    number = {6},
    pages = {35--41},
    title = {Полная анатомическая разметка изображений магнитно-резонансной томографии головного мозга с помощью сопоставления с несколькими атласами},
    year = {2016}
}

Bibtex:

@inproceedings{erofeev15vybor,
    author = {Ерофеев, М. В.},
    booktitle = {XXII Международная конференция студентов, аспирантов и молодых ученых "Ломоносов-2015"},
    language = {russian},
    pages = {14--17},
    publisher = {Издательский отдел факультета вычислительной математики и кибернетики МГУ имени М.В. Ломоносова Москва},
    title = {Выбор метрики для оценки качества алгоритмов матирования видео},
    year = {2015}
}

Bibtex:

@inproceedings{gringauz15algorithm,
    author = {Gringauz, Alexander and Shalnov, Eugeniy and Konushin, Anton},
    booktitle = {Proceedings of the 25th International Conference on Computer Graphics and Vision GraphiCon'2015},
    language = {english},
    pages = {222--226},
    series = {GraphiCon},
    title = {An algorithm for estimating locations of people in world coordinates and its application for tracking},
    year = {2015}
}

Bibtex:

@article{shalnov15improvement,
    address = {Road Town, United Kingdom},
    author = {Shalnov, Eugeniy and Konushin, Vadim and Konushin, Anton},
    doi = {10.1134/S1054661815030220},
    issn = {1555-6212; 1054-6618},
    journal = {Pattern Recognition and Image Analysis: Advances in Mathematical Theory and Applications},
    language = {english},
    pages = {532--540},
    publisher = {Road Town, United Kingdom},
    title = {An improvement on an MCMC-based video tracking algorithm},
    volume = {25},
    year = {2015}
}

Abstract:

We describe a method for the human pose estimation in a video sequence. We propose a new mathematical model of a human pose in a video sequence, which incorporates motion and pose parameters. We show that the model of (Park and Ramanan, 2011) is a particular case of our model. We introduce a framework to infer an approximation of the optimal value in the proposed model. We use an exact algorithm of motion parameters estimation to reduce complexity of inference. Our approach outperforms results of (Park and Ramanan, 2011) in the most complicated video sequences.

Bibtex:

@inproceedings{shalnov15human,
    author = {Shalnov, Evgeny and Konushin, Anton},
    booktitle = {Proceedings of the 5th International Workshop on Image Mining. Theory and Applications},
    language = {english},
    pages = {71--79},
    title = {Human Pose Estimation in Video via MCMC Sampling},
    year = {2015}
}

Bibtex:

@article{kukharenko15simultaneous,
    address = {Road Town, United Kingdom},
    author = {Kukharenko, A. I. and Konushin, A. S.},
    doi = {10.1134/S1054661815030128},
    issn = {1555-6212; 1054-6618},
    journal = {Pattern Recognition and Image Analysis: Advances in Mathematical Theory and Applications},
    language = {english},
    number = {3},
    pages = {461--465},
    publisher = {Road Town, United Kingdom},
    title = {Simultaneous classification of several features of a person’s appearance using a deep convolutional neural network},
    volume = {25},
    year = {2015}
}

Bibtex:

@article{filippov15podschet,
    author = {Филиппов, И. В. and Кононов, В. А. and Конушин, В. С. and Конушин, А. С.},
    issn = {0236-235X},
    journal = {Программные продукты и системы},
    language = {russian},
    number = {1},
    pages = {121--126},
    title = {Подсчет количества людей в видеопоследовательности на основе детектора головы человека},
    year = {2015}
}

Abstract:

Предложен новый алгоритм последовательного определения пола и возраста человека по изображению лица на основе двухэтапной схемы восстановления регрессии с помощью метода опорных векторов. В качестве признакового описания изображения используются только наиболее значимые локальные бинарные шаблоны. Для повышения точности половой классификации применяется процедура бутстраппинга – обучения на трудных примерах, а для повышения точности определения возраста – "плавающие" границы возрастных диапазонов.

Bibtex:

@article{rybintsev15posledovatelnaja,
    author = {Рыбинцев, А. В. and Конушин, В. С. and Конушин, А. С.},
    doi = {10.18287/0134-2452-2015-39-5-762-769},
    issn = {0134-2452},
    journal = {Компьютерная оптика},
    language = {russian},
    number = {5},
    pages = {762--769},
    title = {Последовательная половая и возрастная классификация людей по изображению лица на основе ранжированных локальных бинарных шаблонов},
    volume = {39},
    year = {2015}
}

Bibtex:

@article{gitman15delivering,
    author = {Gitman, Yury and Bal, Can and Erofeev, Mikhail and Jain, Ankit and Matyunin, Sergey and Lee, Kyoung-Rok and Voronov, Alexander and Juang, Jason and Vatolin, Dmitriy and Nguyen, Truong},
    journal = {Intel Technology Journal},
    language = {english},
    number = {1},
    pages = {162--200},
    title = {Delivering Enhanced 3D Video},
    volume = {19},
    year = {2015}
}

Bibtex:

@inproceedings{dolganov15detection,
    author = {Dolganov, Stanislav and Erofeev, Mikhail and Vatolin, Dmitriy and Gitman, Yury},
    booktitle = {2015 International Conference on 3D Imaging, IC3D 2015},
    doi = {10.1109/IC3D.2015.7391839},
    language = {english},
    pages = {1--6},
    publisher = {Liege, Belgium},
    title = {Detection of stuck-to-background objects in converted S3D movies},
    year = {2015}
}

Abstract:

Despite recent progress in the field of video matting, neither public data sets nor even a generally accepted method of measuring quality has yet emerged. In this paper we present an online benchmark for video-matting methods. Using chroma keying and a reflection-aware stop-motion capturing procedure, we prepared 12 test sequences. Then, using subjective data, we performed extensive comparative analysis of different quality metrics. The goal of our benchmark is to enable better understanding of current progress in the field of video matting and to aid in developing new methods.

Bibtex:

@inproceedings{erofeev15perceptually,
    author = {Erofeev, Mikhail and Gitman, Yury and Vatolin, Dmitriy and Fedorov, Alexey and Wang, Jue},
    booktitle = {2015 British Machine Vision Conference (BMVC)},
    doi = {10.5244/C.29.99},
    language = {english},
    pages = {99.1--99.12},
    publisher = {Swansea, United Kingdom},
    title = {Perceptually Motivated Benchmark for Video Matting},
    year = {2015}
}

Bibtex:

@inproceedings{zvezdakov15algoritm,
    author = {Звездаков, С. В. and Боков, А. А. and Ватолин, Д. С.},
    booktitle = {VII Международная ежегодная научно-практическая конференция "Запись и воспроизведение объемных изображений в кинематографе и других отраслях"},
    language = {russian},
    pages = {238--248},
    title = {Алгоритм определения способа производства планов стереофильмов},
    year = {2015}
}

Bibtex:

@inproceedings{zachesov15ispolzovanie,
    author = {Зачесов, А. and Ерофеев, М. and Ватолин, Д.},
    booktitle = {Новые информационные технологии в автоматизированных системах: материалы научно-практического семинара},
    language = {russian},
    publisher = {М.: МИЭМ НИУ ВШЭ},
    title = {Использование карт глубины при восстановлении фона в видеопоследовательностях},
    year = {2015}
}

Bibtex:

@article{erofeev15metodika,
    author = {Ерофеев, М. В. and Гитман, Ю. А. and Ватолин, Д. С. and Федоров, А. А.},
    issn = {1684-2634},
    journal = {Цифровая обработка сигналов},
    language = {russian},
    number = {3},
    pages = {53--59},
    title = {Методика объективного сравнения алгоритмов матирования видео},
    year = {2015}
}

Bibtex:

@inproceedings{bokov15obnaruzhenie,
    author = {Боков, А. А. and Ватолин, Д. С.},
    booktitle = {VII Международная ежегодная научно-практическая конференция "Запись и воспроизведение объемных изображений в кинематографе и других отраслях"},
    language = {russian},
    pages = {223--237},
    title = {Обнаружение артефактов конвертации фильмов в стереоформат},
    year = {2015}
}

Bibtex:

@article{ljudvichenko15obnaruzhenie,
    author = {Людвиченко, В. А. and Лаврушкин, С. В. and Янушковский, В. А. and Ватолин, Д. С.},
    issn = {1991-3400},
    journal = {Мир техники кино},
    language = {russian},
    number = {35},
    pages = {10--18},
    title = {Обнаружение временного сдвига между ракурсами и перепутанного порядка ракурсов в стереофильмах},
    year = {2015}
}

Bibtex:

@article{vatolin15tendentsii,
    author = {Ватолин, Д. С. and Боков, А. А. and Федоров, А. А.},
    issn = {1991-3400},
    journal = {Мир техники кино},
    language = {russian},
    number = {3},
    pages = {17--28},
    title = {Тенденции изменения технического качества стереокино - 5 лет после "Аватара"},
    volume = {37},
    year = {2015}
}

Bibtex:

@inproceedings{vatolin15tendentsiiizmenenija,
    author = {Ватолин, Д. С. and Боков, А. А. and Федоров, А. А.},
    booktitle = {Запись и воспроизведение объемных изображений в кинематографе и других областях: VII Международная научно-практическая конференция},
    language = {russian},
    pages = {68--86},
    publisher = {ВГИК Москва},
    title = {Тенденции изменения технического качества стереокино: 5 лет после "Аватара"},
    year = {2015}
}

Bibtex:

@inproceedings{frolov15cheap,
    author = {Frolov, V. A. and Galaktionov, V. A.},
    booktitle = {Conference proceedings of 25-th International Conference on Computer Graphics and Vision},
    language = {english},
    pages = {28--33},
    publisher = {Институт физико-технической информатики Протвино},
    title = {Cheap and simple in-place block based path regeneration},
    year = {2015}
}

Bibtex:

@article{gruzdev15practical,
    address = {Road Town, United Kingdom},
    author = {Gruzdev, A. M. and Frolov, V. A. and Ignatenko, A. V.},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {5},
    pages = {253--257},
    publisher = {Road Town, United Kingdom},
    title = {Practical Approach to the fast Monte-Carlo Ray Tracing},
    volume = {41},
    year = {2015}
}

Abstract:

Монте-Карло Трассировка Путей является цетральным алгоритмом расчёта освещенности, вокруг ко- торого строятся более современные методы (такие как BDPT, MLT, ERPT, VCM и другие). Одна из основных проблем, стоящих на пути к реализации эффективной трассировки путей на GPU – малая за- грузка GPU вычислениями вследствие сильно различной глубины трассировки: небольшое число потоков трассируют пути на большой глубине, в то время как остальные потоки простаивают. Обычно для решения этой проблемы используется техника, называемая регенерацией путей. Мы предлагаем новый подход к реализации регенерации путей, названный нами ¾блочной регенерацией по месту{\`A}. По сравнению с предыдущими подходами наш алгоритм обладает более низкой стоимостью и не перемещает данные лучей в памяти, за счёт чего упрощается его интеграция в более сложные методы (такие как BDPT, MLT, ERPT, VCM и другие) или фотонные карты. Мы протестировали наш алгоритм для Монте-Карло трассировки путей, используя CUDA и OpenCL.

Bibtex:

@inproceedings{frolov15prostoj,
    author = {Фролов, В. А. and Галактионов, В. А.},
    booktitle = {Труды 25-й Международной конференции по компьютерной графике и зрению ГрафиКон'2015},
    language = {russian},
    pages = {22--25},
    publisher = {Институт физико-технической информатики Протвино},
    series = {GraphiCon},
    title = {Простой и недорогой подход к реализации регенерации путей},
    year = {2015}
}

Projects:

Bibtex:

@inproceedings{senyukova15brain,
    author = {Senyukova, Olga and Zobnin, Denis},
    booktitle = {Pattern Recognition (37th German Conference, GCPR 2015, Aachen, Germany, October 7–10, 2015, Proceedings)},
    doi = {10.1007/978-3-319-24947-6\_32},
    language = {english},
    pages = {390--399},
    publisher = {Springer International Publishing Heidelberg, Germany},
    series = {Lecture Notes in Computer Science},
    title = {Fast Brain MRI Registration with Automatic Landmark Detection Using a Single Template Image},
    volume = {9358},
    year = {2015}
}

Bibtex:

@inproceedings{senyukova15generic,
    author = {Senyukova, Olga and Gavrishchaka, Valeriy},
    booktitle = {BIT's 7th Annual International Congress of Cardiology-2015 (Shanghai, China)},
    language = {english},
    pages = {176},
    title = {Generic Multi-Complexity Representation of Cardiodynamics: From Early Detection of Emerging Abnormalities to Personalized Treatment Optimization},
    year = {2015}
}

Bibtex:

@inproceedings{gavrishchaka15multicomplexity,
    author = {Gavrishchaka, Valeriy and Senyukova, Olga and Davis, Kristina},
    booktitle = {Signal and Image Analysis for Biomedical and Life Sciences},
    doi = {10.1007/978-3-319-10984-8\_6},
    language = {english},
    pages = {107--126},
    publisher = {Springer International Publishing Switzerland},
    series = {Advances in Experimental Medicine and Biology},
    title = {Multi-Complexity Ensemble Measures for Gait Time Series Analysis: Application to Diagnostics, Monitoring and Biometrics},
    volume = {823},
    year = {2015}
}

Bibtex:

@inproceedings{senjukova15algoritm,
    author = {Сенюкова, О. and Зобнин, Д. and Петряйкин, А.},
    booktitle = {XVII Всероссийская научно- техническая конференция "Нейроинформатика-2015" с международным участием: Сборник научных трудов},
    language = {russian},
    pages = {71--80},
    publisher = {НИЯУ МИФИ Москва},
    title = {Алгоритм сопоставления изображений магнитно-резонансной томографии головного мозга человека на основе ключевых точек},
    volume = {2},
    year = {2015}
}

Bibtex:

@inproceedings{zubov15segmentatsija,
    author = {Зубов, А. and Сенюкова, О.},
    booktitle = {Труды 25-й Международной конференции по компьютерной графике и зрению ГрафиКон'2015},
    language = {russian},
    pages = {56--61},
    publisher = {Институт физико-технической информатики Протвино},
    series = {GraphiCon},
    title = {Сегментация изображений магнитно-резонансной томографии головного мозга с помощью сопоставления с несколькими атласами},
    year = {2015}
}

Bibtex:

@inproceedings{nikitin14quality,
    author = {Nikitin, М. and Konushin, A. and Konushin, V.},
    booktitle = {Proceedings of the 24th International Conference on Computer Graphics and Vision GraphiCon'2014},
    language = {english},
    pages = {111--114},
    publisher = {Rostov-on-Don, Russia},
    series = {GraphiCon},
    title = {Face quality assessment for face verification in video},
    year = {2014}
}

Bibtex:

@inproceedings{gringauz14modification,
    author = {Gringauz, A. and Shalnov, E. and Konushin, A.},
    booktitle = {Proceedings of the 24th International Conference on Computer Graphics and Vision GraphiCon'2014},
    language = {english},
    pages = {139--143},
    publisher = {Rostov-on-Don, Russia},
    series = {GraphiCon},
    title = {Modification of the multi-target tracking algorithm based on energy minimization},
    year = {2014}
}

Abstract:

In this paper we present algorithms for automatically detecting issues specific to converted S3D content. When a depth-image-based rendering approach produces a stereoscopic image, the quality of the result depends on both the depth maps and the warping algorithms. The most common problem with converted S3D video is edge-sharpness mismatch. This artifact may appear owing to depth-map blurriness at semitransparent edges: after warping, the object boundary becomes sharper in one view and blurrier in the other, yielding binocular rivalry. To detect this problem we estimate the disparity map, extract boundaries with noticeable differences, and analyze edge-sharpness correspondence between views. We pay additional attention to cases involving a complex background and large occlusions. Another problem is detection of scenes that lack depth volume: we present algorithms for detecting at scenes and scenes with at foreground objects. To identify these problems we analyze the features of the RGB image as well as uniform areas in the depth map. Testing of our algorithms involved examining 10 Blu-ray 3D releases with converted S3D content, including Clash of the Titans, The Avengers, and The Chronicles of Narnia: The Voyage of the Dawn Treader. The algorithms we present enable improved automatic quality assessment during the production stage.

Bibtex:

@inproceedings{bokov14automatic,
    author = {Bokov, Alexander A. and Vatolin, Dmitriy and Zachesov, Anton and Belous, Alexander and Erofeev, Mikhail},
    booktitle = {Proc. SPIE 9011, Stereoscopic Displays and Applications XXV (March 6, 2014)},
    doi = {10.1117/12.2054330},
    language = {english},
    pages = {1--14},
    publisher = {San Francisco, California, USA},
    title = {Automatic detection of artifacts in converted S3D video},
    volume = {901112},
    year = {2014}
}

Abstract:

This research aims to sufficiently increase the quality of visual-attention modeling to enable practical applications. We found that automatic models are significantly worse at predicting attention than even single-observer eye tracking. We propose a semiautomatic approach that requires eye tracking of only one observer and is based on time consistency of the observer’s attention. Our comparisons showed the high objective quality of our proposed approach relative to automatic methods and to the results of single-observer eye tracking with no postprocessing. We demonstrated the practical applicability of our proposed concept to the task of saliency-based video compression.

Bibtex:

@inproceedings{gitman14semiautomatic,
    author = {Gitman, Yury and Erofeev, Mikhail and Vatolin, Dmitriy and Bolshakov, Andrey and Fedorov, Alexey},
    booktitle = {2014 IEEE International Conference on Image Processing (ICIP) (ICIP 2014)},
    doi = {10.1109/ICIP.2014.7025220},
    language = {english},
    pages = {1105--1109},
    publisher = {Paris, France},
    title = {Semiautomatic Visual-Attention Modeling and Its Application to Video Compression},
    year = {2014}
}

Bibtex:

@inproceedings{ljudvichenko14obnaruzhenie,
    author = {Людвиченко, В. А. and Лаврушкин, С. В. and Янушковский, В. А. and Ватолин, Д. С.},
    booktitle = {VI Международная научно-технической конференция "Запись и воспроизведение объёмных изображений в кинематографе и других отраслях"},
    language = {russian},
    publisher = {Москва},
    title = {Обнаружение временного сдвига между ракурсами и перепутанного порядка ракурсов в стереофильмах},
    year = {2014}
}

Abstract:

Рассматривается задача полуавтоматического построения карты глубины для видео. Предложен метод обработки областей наложений объектов для повышения качества карт глубины. Информация о таких областях аккумулируется при последовательной обработке кадров видео, а затем используется для маскирования ненадежных интерполированных областей. Приведено сравнение алгоритмов поиска областей наложений объектов и способов их использования в рассматриваемой задаче.

Bibtex:

@article{matjunin14obrabotka,
    author = {Матюнин, С. Б. and Ватолин, Д. С.},
    issn = {1684-2634},
    journal = {Цифровая обработка сигналов},
    language = {russian},
    pages = {3--23},
    title = {Обработка областей наложений объектов при построении карт глубины для видео},
    year = {2014}
}

Abstract:

В данной статье предлагается эффективная и исключительно простая реализация алгоритма фотонных карт на GPU на основе окто-деревьев с множественными ссылками. Предлагаемый алгоритм построения дерева состоит из 4 шагов, использует линейные структуры данных и только 2 параллельных примитива для построения дерева. Результирующая структура позволяет производить эффективный поиск ближайших фотонов в окто-дереве с заданным радиусом сбора без использования стэка.

Bibtex:

@article{frolov14oktoderevja,
    author = {Фролов, В. А. and Харламов, А. А. and Галактионов, В. А. and Востряков, К. А.},
    doi = {10.1134/S0361768814040033},
    issn = {0132-3474},
    journal = {Программирование},
    language = {russian},
    number = {4},
    pages = {64--73},
    title = {Окто-деревья со множественными ссылками в применении к реализации фотонных карт и кэша освещенности на GPU},
    volume = {40},
    year = {2014}
}

Abstract:

В данной статье представлен сравнительный анализ производительности современных рендер-систем. Нами было проведено сравнение среди 7 наиболее популярными продуктов для 3D Studio Max (Mental Ray, IRay, VRay3, VRayRT3, Corona, Octane, Hydra) на скорость интегрирования освещения на 7 различных сценариях. В нашей работе нас прежде всего интересовало сравнение законченных реализации в целом в промышленном решении в противовес сравнению отдельных алгоритмов в одной и той же системе. Мы постарались ответить на некоторые важные вопросы: что на практике дают последние достижения в области алгоритмов, такие как SPPM [1] и VCM [2]? Есть ли выигрыш от GPU? Смещенные и несмещенные рендер-системы - что и когда лучше? Какие алгоритмы несут наибольшую практическую пользу? На основе проведенного исследования нами были сделаны заключения о современном положении дел в области рендер-систем, преимуществах и недостатках CPU и GPU реализаций, а также о достоинствах и недостатках различных алгоритмов, применяемых в исследуемых системах и возможных векторах развития. Рендер-системы были ранжированы по очкам относительной производительности для каждого из сценариев и в сумме. Наши тестовые сцены и данные находятся в открытом доступе.

Bibtex:

@inproceedings{frolov14sravnitelnyj,
    author = {Фролов, В. А. and Галактионов, В. А. and Трофимов, М. А.},
    booktitle = {Труды 24-й Международной конференции по компьютерной графике и зрению ГрафиКон'2014},
    language = {russian},
    pages = {43--46},
    publisher = {Академия архитектуры и искусств ЮФУ Ростов-на-Дону},
    series = {GraphiCon},
    title = {Сравнительный анализ современных рендер-систем для 3ds Max},
    year = {2014}
}

Bibtex:

@book{senjukova14sbalansirovannye,
    author = {Сенюкова, О. В.},
    isbn = {978-5-89407-528-0; 978-5-317-04873-0},
    language = {russian},
    numpages = {68},
    publisher = {Издательский отдел факультета ВМК МГУ имени М.В. Ломоносова, МАКС Пресс Москва},
    title = {Сбалансированные деревья поиска: Учебно-методическое пособие},
    year = {2014}
}

Projects:

Bibtex:

@article{senyukova14segmentation,
    address = {United States},
    author = {Senyukova, O. V.},
    doi = {10.1016/j.patcog.2014.06.007},
    issn = {0031-3203},
    journal = {Pattern Recognition},
    language = {english},
    number = {12},
    pages = {3881--3889},
    publisher = {United States},
    title = {Segmentation of Blurred Objects by Classification of Isolabel Contours},
    volume = {47},
    year = {2014}
}

Bibtex:

@inproceedings{senyukova14universal,
    address = {HEIDELBERGER PLATZ 3, BERLIN, GERMANY,D-14197},
    author = {Senyukova, O. and Gavrishchaka, V. and Koepke, M.},
    booktitle = {Proceedings of The First International Aizu Conference on Biomedical Informatics and Technology (ACBIT 2013)},
    doi = {10.1007/978-3-642-54121-6\_7},
    language = {english},
    pages = {76--90},
    publisher = {HEIDELBERGER PLATZ 3, BERLIN, GERMANY,D-14197},
    series = {Communications in Computer and Information Science},
    title = {Universal Multi-Complexity Measures for Physiological State Quantification in Intelligent Diagnostics and Monitoring Systems},
    volume = {404},
    year = {2014}
}

Bibtex:

@inproceedings{erofeev13razrabotka,
    author = {Ерофеев, М. В.},
    booktitle = {XX Международная конференция студентов, аспирантов и молодых ученых "Ломоносов-2013"},
    language = {russian},
    pages = {100--102},
    publisher = {МАКС Пресс Москва},
    title = {Разработка стабильного во времени метода матирования объектов переднего плана в видео},
    year = {2013}
}

Abstract:

We present a system for the large-scale automatic traffic signs recognition and mapping and experimentally justify design choices made for different components of the system. Our system works with more than 140 different classes of traffic signs and does not require labor-intensive labelling of a large amount of training data due to the training on synthetically generated images. We evaluated our system on the large dataset of Russian traffic signs and made this dataset publically available to encourage future comparison

Projects:

Bibtex:

@article{chigorin13system,
    author = {Chigorin, A. and Konushin, A.},
    doi = {10.5194/isprsannals-II-3-W3-13-2013},
    journal = {ISPRS Annals of Photogrammetry, Remote Sensing and Spatial Information Sciences},
    language = {english},
    number = {II-3/W3},
    pages = {13--17},
    title = {A system for large-scale automatic traffic sign recognition and mapping},
    year = {2013}
}

Bibtex:

@inproceedings{sindeev13alphaflow,
    author = {Sindeev, Mikhail and Konushin, Anton and Rother, Carsten},
    booktitle = {Computer Vision – ACCV 2012},
    doi = {10.1007/978-3-642-37431-9\_34},
    language = {english},
    pages = {438--452},
    publisher = {Springer Berlin Heidelberg},
    series = {Lecture Notes in Computer Science},
    title = {Alpha-Flow for Video Matting},
    volume = {7726},
    year = {2013}
}

Projects:

Bibtex:

@inproceedings{moiseev13evaluation,
    author = {Moiseev, Boris and Konev, Artem and Chigorin, Alexander and Konushin, Anton},
    booktitle = {Advanced Concepts for Intelligent Vision Systems},
    doi = {10.1007/978-3-319-02895-8\_52},
    language = {english},
    pages = {576--583},
    publisher = {Springer},
    series = {Lecture Notes in Computer Science},
    title = {Evaluation of Traffic Sign Recognition Methods Trained on Synthetically Generated Data},
    volume = {8192},
    year = {2013}
}

Abstract:

In the article we describe a problem of a fully automatic people tracking in surveillance video recorded by a stable camera. We propose a development of Benfold et al tracking-by-detection algorithm [1]. We introduce a filtering of person detector results and a modified method of tracklet’s position estimation. Also we use the Flock of features and the normalized cross-correlation, as visual tracking algorithms, to construct tracklets. Our experiments reveal that all proposed modifications allow producing more robust trajectories than trajectories produced by basic algorithm.

Bibtex:

@inproceedings{shalnov13improvement,
    author = {Shalnov, E. V. and Konushin, V. S. and Konushin, A. S.},
    booktitle = {11th International Conference on Pattern Recognition and Image Analysis: New Information Technologies (PRIA-11-2003). Samara, September 23-28, 2013. Conference Proceedings},
    language = {english},
    pages = {727--730},
    publisher = {IPSI RAS Samara},
    title = {Improvement of MCMC-based video tracking algorithm},
    volume = {2},
    year = {2013}
}

Bibtex:

@inproceedings{kuharenko13simultaneous,
    author = {Kuharenko, A. I. and Konushin, A. S.},
    booktitle = {11th International Conference on Pattern Recognition and Image Analysis: New Information Technologies (PRIA-11-2003). Samara, September 23-28, 2013. Conference Proceedings},
    language = {english},
    pages = {623--626},
    publisher = {IPSI RAS Samara},
    title = {Simultaneous facial attribute classification with convolutional neural networks},
    volume = {2},
    year = {2013}
}

Bibtex:

@article{rybintsev13vozrastnaja,
    address = {М.},
    author = {Рыбинцев, А. В. and Лукина, Т. М. and Конушин, В. С. and Конушин, А. С.},
    issn = {0869-6527},
    journal = {Системы и средства информатики},
    language = {russian},
    number = {2},
    pages = {48--59},
    publisher = {М.},
    title = {Возрастная классификация людей по изображению лица на основе метода ранжирования и локальных бинарных шаблонов},
    volume = {23},
    year = {2013}
}

Bibtex:

@article{konushin13klassifikatsija,
    address = {М.},
    author = {Конушин, В. and Лукина, Т. М. and Кухаренко, А. И. and Конушин, А. С.},
    issn = {0869-6527},
    journal = {Системы и средства информатики},
    language = {russian},
    number = {2},
    pages = {34--45},
    publisher = {М.},
    title = {Классификация людей по изображению лица на основе сравнительных признаков внешности},
    volume = {23},
    year = {2013}
}

Bibtex:

@inproceedings{kuharenko13opredelenie,
    author = {Кухаренко, А. И. and Конушин, А. С.},
    booktitle = {Сборник тезисов научно-технической конференции "Техническое зрение в системах управления"},
    language = {russian},
    pages = {146--148},
    title = {Определение пола человека с помощью глубоких нейронных сетей},
    year = {2013}
}

Bibtex:

@inproceedings{kononov13opredelenie,
    author = {Кононов, В. А. and Конушин, В. С. and Конушин, А. С.},
    booktitle = {Сборник тезисов научно-технической конференции "Техническое зрение в системах управления"},
    language = {russian},
    pages = {148--150},
    title = {Определение типа объекта в видео с камер наблюдения},
    year = {2013}
}

Bibtex:

@inproceedings{kononov13opredelenieobekta,
    author = {Кононов, В. А. and Конушин, В. С. and Конушин, А. С.},
    booktitle = {Сборник тезисов научно-технической конференции "Техническое зрение в системах управления"},
    language = {russian},
    pages = {148--150},
    title = {Определение типа объекта в видео с камер наблюдения},
    year = {2013}
}

Bibtex:

@article{kononov13opredelenietipov,
    author = {Кононов, В. А. and Конушин, А. С.},
    issn = {1810-7206},
    journal = {Вестник компьютерных и информационных технологий},
    language = {russian},
    number = {10},
    pages = {20--25},
    title = {Определение типов объектов в видеопотоке с камеры наблюдения на основе покадровой классификации},
    year = {2013}
}

Bibtex:

@article{chigorin13sistema,
    author = {Чигорин, А. A. and Конушин, А. C.},
    issn = {0236-235X},
    journal = {Программные продукты и системы},
    language = {russian},
    number = {3},
    pages = {288--291},
    title = {Система автоматического картографирования знаков дорожного движения},
    year = {2013}
}

Bibtex:

@inproceedings{batanov13soprovozhdenie,
    author = {Батанов, П. and Кононов, В. А. and Конушин, А. С.},
    booktitle = {23-я Международная конференция по компьютерной графике и зрению: Владивосток, Институт автоматики и процессов управления ДВОРАН: Труды конференции},
    language = {russian},
    pages = {333--336},
    publisher = {"Дальнаука" ДВО РАН Владивосток, Россия},
    title = {Сопровождение людей в системах многокамерного видеонаблюдения для спортивных игр},
    year = {2013}
}

Bibtex:

@article{chigorin13eksperimenty,
    author = {Чигорин, А. A. and Конушин, А. C.},
    issn = {1994-0408},
    journal = {Наука и образование (МГТУ им. Н.Э. Баумана) (электронный журнал)},
    language = {russian},
    number = {8},
    pages = {315--324},
    title = {Эксперименты с обучением методов распознавания дорожных знаков на синтетических данных},
    year = {2013}
}

Abstract:

We propose a method of 3D video compression based on 2D+depth representation. We use correlation between 2D video and depth map to reconstruct highly compressed depth map while decoding. Depth map is compressed with reduced spatial and temporal resolution. On decoding stage the algorithm restore original resolution using information from 2D video. We evaluated the influence of key frames’ resolution, compression ratio and density on the performance of the algorithm. The proposed technique was compared to depth map compression using H.264 in compression pipeline for stereo video.

Bibtex:

@inproceedings{matyunin13video,
    author = {Matyunin, Sergey and Vatolin, Dmitriy},
    booktitle = {Multimedia Communications, Services and Security},
    doi = {10.1007/978-3-642-38559-9\_14},
    language = {english},
    pages = {153--166},
    publisher = {Springer Berlin Heidelberg},
    title = {3D Video Compression Using Depth Map Propagation},
    year = {2013}
}

Abstract:

Creating and processing stereoscopic video imposes additional quality requirements related to view synchronization. In this work we propose a set of algorithms for detecting typical stereoscopic-video problems, which appear owing to imprecise setup of capture equipment or incorrect postprocessing. We developed a methodology for analyzing the quality of S3D motion pictures and for revealing their most problematic scenes. We then processed 10 modern stereo films, including Avatar, Resident Evil: Afterlife and Hugo, and analyzed changes in S3D-film quality over the years. This work presents real examples of common artifacts (color and sharpness mismatch, vertical disparity and excessive horizontal disparity) in the motion pictures we processed, as well as possible solutions for each problem. Our results enable improved quality assessment during the filming and postproduction stages.

Bibtex:

@inproceedings{voronov13methodology,
    address = {Bellingham, WA, United States},
    author = {Voronov, Alexander and Vatolin, Dmitriy and Sumin, Denis and Napadovsky, Vyacheslav and Borisov, Alexey},
    booktitle = {Proc. SPIE 8648, Stereoscopic Displays and Applications XXIV},
    doi = {10.1117/12.2008485},
    language = {english},
    pages = {864810-1--864810-14},
    publisher = {Bellingham, WA, United States},
    title = {Methodology for stereoscopic motion-picture quality assessment},
    volume = {8648},
    year = {2013}
}

Bibtex:

@article{bolshakov13voprosy,
    author = {Большаков, А. and Федоров, А. and Ватолин, Д.},
    issn = {1991-3400},
    journal = {Мир техники кино},
    language = {russian},
    number = {28},
    pages = {10--16},
    title = {Вопросы построения базы записей движений глаз для анализа восприятия кинофильмов},
    year = {2013}
}

Bibtex:

@inproceedings{bolshakov13voprosypostroenija,
    author = {Большаков, А. С. and Федоров, А. А. and Ватолин, Д. С.},
    booktitle = {Международная ежегодная научно-техническая конференция "Запись и воспроизведение объемных изображений в кинематографе и других областях"},
    language = {russian},
    pages = {54--70},
    title = {Вопросы построения базы записей движений глаз для анализа восприятия кинофильмов},
    year = {2013}
}

Bibtex:

@inproceedings{bolshakov13sozdanie,
    author = {Большаков, А. С. and Федоров, А. А. and Ватолин, Д. С.},
    booktitle = {Конференция-школа молодых ученых и специалистов "Информационные технологии и системы"},
    language = {russian},
    pages = {1--6},
    title = {Создание базы записей движений глаз при просмотре видеоряда для верификации модели внимания},
    volume = {37},
    year = {2013}
}

Abstract:

This work presents an approach to integrating irradiance caching (IC) technique in a complete GPU photorealistic renderer. This work proposes a GPU friendly IC solution, where performance critical parts of an irradiance cache algorithm are done completely on the GPU. The modified algorithm for the GPU is different from a traditional implementation in 2 ways. The first distinction is a predictive nature of our algorithm that allows us to insert a large record set at once instead of inserting records one by one, as in traditional approaches. The second distinction is a new heuristic for validity radius computations. We also consider some low-level details and provide performance analysis of our solution.

Bibtex:

@inproceedings{frolov13implementing,
    author = {Frolov, Vladimir and Vostryakov, Konstantin and Kharlamov, Alexander and Galaktionov, Vladimir},
    booktitle = {Transactions on Computational Science XIX: Special Issue on Computer Graphics},
    doi = {10.1007/978-3-642-39759-2\_2},
    language = {english},
    pages = {17--32},
    publisher = {Springer Berlin Heidelberg},
    series = {Lecture Notes in Computer Science},
    title = {Implementing Irradiance Cache in a GPU Realistic Renderer},
    volume = {7870},
    year = {2013}
}

Abstract:

In this paper we propose a new approach to progressive video rendering. We apply a multidimensional filtering to samples with additional information about a scene along with color. Our filter takes noisy output of path tracing with a low amount of paths per pixel and yields the resulting movie quicker than tracing a large amount of rays. Our approach adds only a linear member O(N) to a path tracing computing complexity (where N is amount of pixels). With the help of progressive scheme the resulting video quality is improving after the every iteration. Using our approach we can get a video of a good quality even computing 10 paths per pixel.

Bibtex:

@inproceedings{gruzdev13multidimetial,
    author = {Gruzdev, Alexei and Frolov, Vladimir and Vostryakov, Konstantin and Ignatenko, Alexei},
    booktitle = {Proceedings of the 23rd International Conference on Computer Graphics and Vision GraphiCon'2013},
    language = {english},
    pages = {75--78},
    publisher = {IACP FEB RAS, Vladivostok, Russia Vladivostok, Russia},
    series = {GraphiCon},
    title = {Multidimetial Filtering in application to Progressive Video Rendering},
    year = {2013}
}

Bibtex:

@inproceedings{gavrishchaka13multicomplexity,
    author = {Gavrishchaka, V. and Davis, K. and Senyukova, O.},
    booktitle = {2013 INTERNATIONAL SYMPOSIUM ON COMPUTATIONAL MODELS FOR LIFE SCIENCES (CMLS-13)},
    language = {english},
    pages = {47--56},
    series = {AIP Conference Proceedings},
    title = {Multi-complexity measures for early detection and monitoring of neurological abnormalities from gait time series},
    volume = {1559},
    year = {2013}
}

Bibtex:

@inproceedings{gavrishchaka13robust,
    address = {Germany},
    author = {Gavrishchaka, ValeriyV and Senyukova, OlgaV},
    booktitle = {Knowledge-Based Systems in Biomedicine and Computational Life Science},
    language = {english},
    pages = {137--153},
    publisher = {Germany},
    series = {Studies in Computational Intelligence},
    title = {Robust Algorithmic Detection of Cardiac Pathologies from Short Periods of RR Data},
    volume = {450},
    year = {2013}
}

Bibtex:

@inproceedings{erofeev12korrektnaja,
    author = {Ерофеев, М. В.},
    booktitle = {XIX Международная конференция студентов, аспирантов и молодых ученых "Ломоносов-2012"},
    language = {russian},
    pages = {22--23},
    publisher = {Издательский отдел факультета вычислительной математики и кибернетики Москва},
    series = {Секция "Вычислительная математика и кибернетика"},
    title = {Корректная обработка полупрозрачных границ при построении стереоизображения},
    year = {2012}
}

Bibtex:

@inproceedings{chigorin12method,
    author = {Chigorin, A. and Krivovyaz, G. and Velizhev, A. and Konushin, A.},
    booktitle = {14th International Conference Digital Signal Processing and its Applications},
    language = {english},
    pages = {316--319},
    title = {A method for traffic sign detection in an image with learning from synthetic data},
    volume = {2},
    year = {2012}
}

Bibtex:

@inproceedings{konushin12simile,
    author = {Konushin, V. and Lukina, T. and Kuharenko, A. and Konushin, A.},
    booktitle = {Proceedings of the 22nd International Conference on Computer Graphics and Vision GraphiCon'2012},
    language = {english},
    pages = {108--112},
    publisher = {Moscow, Russia},
    series = {GraphiCon},
    title = {Simile Classifiers for Face Classification},
    year = {2012}
}

Bibtex:

@article{zejnalov12beskontaktnye,
    author = {Зейналов, Р. Ш. and Конушин, А. С. and Якубенко, А. А.},
    journal = {НОВЫЕ ИНФОРМАЦИОННЫЕ ТЕХНОЛОГИИ В АВТОМАТИЗИРОВАННЫХ СИСТЕМАХ},
    language = {russian},
    number = {15},
    pages = {51--58},
    title = {Бесконтактные измерения характеристик движения объекта с использованием инфракрасных маркеров},
    year = {2012}
}

Bibtex:

@inproceedings{chetverikov12vydelenie,
    author = {Четвериков, Н. and Конушин, А.},
    booktitle = {Труды 22-й Международной конференции по компьютерной графике и зрению ГрафиКон'2012},
    language = {russian},
    pages = {262--265},
    publisher = {Московский государственный университет им. М.В.Ломоносова Москва},
    series = {GraphiCon},
    title = {Выделение объектов в видеопотоке на основе разрезов графов},
    year = {2012}
}

Bibtex:

@inproceedings{shalnov12kaskadnyj,
    author = {Шальнов, Е. В. and Кононов, В. А. and Конушин, В. С. and Конушин, А. С.},
    booktitle = {Сборник трудов научно-технической конференции "Техническое зрение в системах управления 2012"},
    language = {russian},
    pages = {151--157},
    publisher = {Москва},
    title = {Каскадный алгоритм для отслеживания лица человека в видеопотоке},
    year = {2012}
}

Bibtex:

@inproceedings{sindeyev12mnogokadrovyj,
    author = {Sindeyev, М. and Konushin, А. and Rother, C.},
    booktitle = {Труды 22-й Международной конференции по компьютерной графике и зрению ГрафиКон'2012},
    language = {russian},
    pages = {288--291},
    publisher = {Московский государственный университет им. М.В.Ломоносова Москва},
    series = {GraphiCon},
    title = {Многокадровый оптический поток на основе траекторий},
    year = {2012}
}

Bibtex:

@inproceedings{rybintsev12modifikatsija,
    author = {Рыбинцев, А. and Лукина, Т. and Конушин, В. and Конушин, А.},
    booktitle = {Труды 22-й Международной конференции по компьютерной графике и зрению ГрафиКон'2012},
    language = {russian},
    pages = {133--137},
    publisher = {Московский государственный университет им. М.В.Ломоносова Москва},
    series = {GraphiCon},
    title = {Модификация алгоритма классификации возраста людей по изображению лица на основе метода ранжирования},
    year = {2012}
}

Bibtex:

@inproceedings{chigorin12raspoznavanie,
    author = {Чигорин, А. and Конев, А. and Кривовязь, Г. and Велижев, А. and Конушин, А.},
    booktitle = {Сборник трудов научно-технической конференции "Техническое зрение в системах управления 2012"},
    language = {russian},
    pages = {165--169},
    publisher = {Москва},
    title = {Распознавание знаков дорожного движения на изображениях с обучением на синтетических данных},
    year = {2012}
}

Bibtex:

@inproceedings{zejnalov12rekonstruktsija,
    author = {Зейналов, Р. and Конушин, A.},
    booktitle = {Труды 22-й Международной конференции по компьютерной графике и зрению ГрафиКон'2012},
    language = {russian},
    pages = {292--295},
    publisher = {Московский государственный университет им. М.В.Ломоносова Москва},
    series = {GraphiCon},
    title = {Реконструкция траектории движения и формы объекта с использованием инфракрасных маркеров и стереопары},
    year = {2012}
}

Bibtex:

@article{konushin12semanticheskaja,
    author = {Конушин, АС and Шаповалов, РВ and Велижев, АБ and Баринова, ОВ},
    issn = {0236-235X},
    journal = {Программные продукты и системы},
    language = {russian},
    number = {1},
    pages = {47--52},
    title = {Семантическая сегментация данных лазерного сканирования},
    year = {2012}
}

Abstract:

Automatic analysis of stereo-video quality plays an important role in the process of capturing, converting and editing video in 3D format. Although several low-level stereo-video quality metrics have been proposed, mane more-challenging problems of high-level stereo-video analysis, such as left-right channel swap detection, are still practically unresearched. The visual result of a channel swap is very disconcerting, but it is not always obvious even to a human observer what is wrong in such a sequence. In this paper we present a fully automatic algorithm for left-right channel swap detection. Experimental results for real video sequences demonstrate the effectiveness of the proposed technique.

Bibtex:

@inproceedings{akimov12automatic,
    author = {Akimov, D. and Shestov, A. and Voronov, A. and Vatolin, D.},
    booktitle = {International Conference on 3D Imaging},
    doi = {doi:10.1109/IC3D.2012.6615126},
    language = {english},
    pages = {1--6},
    title = {Automatic Left-Right Channel Swap Detection},
    year = {2012}
}

Bibtex:

@inproceedings{zachesov12background,
    author = {Zachesov, Anton and Vatolin, Dmitry and Smirnov, Maxim},
    booktitle = {Proceedings of the 22nd International Conference on Computer Graphics and Vision GraphiCon'2012},
    language = {english},
    pages = {25--28},
    publisher = {Moscow, Russia},
    series = {GraphiCon},
    title = {Background mosaic reconstruction},
    year = {2012}
}

Abstract:

An algorithm for automatic swapped views detection is proposed. No analogues in literature were found for this problem solution. It is based on occlusion detection and motion vectors histogram. The algorithm was tested on 780 frames from 13 movies. The recall{\textbackslash}precision diagrams were constructed using two parameters. The complexity is estimated. The drawbacks are analyzed and further directions are proposed.

Bibtex:

@inproceedings{shestov12detection,
    author = {Shestov, Alexey and Voronov, Alexander and Vatolin, Dmitriy},
    booktitle = {Proceedings of the 22nd International Conference on Computer Graphics and Vision GraphiCon'2012},
    language = {english},
    pages = {23--27},
    publisher = {Moscow, Russia},
    series = {GraphiCon},
    title = {Detection of swapped views in stereo image},
    year = {2012}
}

Bibtex:

@inproceedings{akimov12occlusion,
    author = {Akimov, D. and Shestov, A. and Voronov, A. and Vatolin, D.},
    booktitle = {International Conference on 3D Imaging},
    doi = {doi:10.1109/IC3D.2012.6615127},
    language = {english},
    pages = {1--8},
    title = {Occlusion Refinement for Stereo Video Using Optical Flow},
    year = {2012}
}

Bibtex:

@inproceedings{voronov12system,
    author = {Voronov, A. and Borisov, A. and Vatolin, D.},
    booktitle = {International Workshop on Video Processing and Quality Metrics for Consumer Electronic (VPQM-2012)},
    language = {english},
    pages = {138--143},
    title = {System for automatic detection of distorted scenes in stereo video},
    year = {2012}
}

Bibtex:

@inproceedings{vatolin12testing,
    author = {Vatolin, Dmitriy and Erofeev, Mikhail and Zachesov, Anton and Sumin, Denis and Akimov, Dmitriy and Fedorov, Alexey},
    booktitle = {International Workshop on Video Processing and Quality Metrics for Consumer Electronic (VPQM-2012)},
    language = {english},
    title = {Testing Methods for 3D Content Viewing Devices},
    year = {2012}
}

Bibtex:

@inproceedings{erofeev12toward,
    author = {Erofeev, M. and Vatolin, D. and Voronov, A. and Fedorov, A.},
    booktitle = {International Conference on 3D Imaging},
    doi = {doi:10.1109/IC3D.2012.6615120},
    language = {english},
    pages = {1--6},
    title = {Toward an Objective Stereo-Video Quality Metric: Depth Perception of Textured Areas},
    year = {2012}
}

Bibtex:

@inproceedings{voronov12towards,
    author = {Voronov, A. and Vatolin, D. and Sumin, D. and Napadovsky, V. and Borisov, A.},
    booktitle = {International Conference on 3D Imaging},
    doi = {doi:10.1109/IC3D.2012.6615121},
    language = {english},
    pages = {1--6},
    title = {Towards Automatic Stereo-video Quality Assessment and Detection of Color and Sharpness Mismatch},
    year = {2012}
}

Bibtex:

@inproceedings{vatolin12issledovanie,
    author = {Ватолин, Д. С. and Воронов, А. А. and Нападовский, В. В. and Борисов, А. В.},
    booktitle = {Международная ежегодная научно-техническая конференция "Запись и воспроизведение объемных изображений в кинематографе и других областях"},
    language = {russian},
    pages = {190--203},
    title = {Исследование артефактов в стереокино и примеры анализа кинокартин},
    year = {2012}
}

Bibtex:

@article{sumin12poluavtomaticheskoe,
    author = {Сумин, Д. А. and Федоров, А. А. and Ватолин, Д. С.},
    issn = {1991-3400},
    journal = {Мир техники кино},
    language = {russian},
    number = {23},
    pages = {5--7},
    title = {Полуавтоматическое определение характеристик устройств отображения объёмного видео},
    year = {2012}
}

Bibtex:

@inproceedings{sumin12poluavtomaticheskoeopredelenie,
    author = {Сумин, Д. А. and Фёдоров, А. А. and Ватолин, Д. С.},
    booktitle = {Международная ежегодная научно-техническая конференция "Запись и воспроизведение объемных изображений в кинематографе и других областях"},
    language = {russian},
    pages = {210--215},
    title = {Полуавтоматическое определение характеристик устройств отображения объёмного видео},
    year = {2012}
}

Bibtex:

@article{zachesov12postroenie,
    author = {Зачёсов, АА and Колиниченко, МБ and Ватолин, ДС},
    issn = {1991-3400},
    journal = {Мир техники кино},
    language = {russian},
    number = {23},
    pages = {19--21},
    title = {Построение карты глубины и многоракурсного видео по стереопаре},
    year = {2012}
}

Bibtex:

@article{matjunin12szhatie,
    author = {Матюнин, Сергей and Ватолин, Дмитрий},
    issn = {1991-3400},
    journal = {Мир техники кино},
    language = {russian},
    number = {25},
    pages = {2--9},
    title = {Сжатие карт глубины при передаче стереовидео для цифрового телевизионного вещания},
    year = {2012}
}

Bibtex:

@inproceedings{matjunin12szhatieglubiny,
    author = {Матюнин, С. Б. and Ватолин, Д. С.},
    booktitle = {Международная ежегодная научно-техническая конференция "Запись и воспроизведение объемных изображений в кинематографе и других областях"},
    language = {russian},
    pages = {232--250},
    title = {Сжатие карт глубины при передаче стереовидео для цифрового телевизионного вещания},
    year = {2012}
}

Bibtex:

@article{matjunin12szhatiemnogorakursnogo,
    author = {Матюнин, СБ and Ватолин, ДС},
    issn = {1991-3400},
    journal = {Мир техники кино},
    language = {russian},
    number = {23},
    pages = {8--10},
    title = {Сжатие многоракурсного видео на основе межкадрового распространения карт глубины},
    year = {2012}
}

Abstract:

Данная книга представляет собой подробное практическое руководство по разработке приложений с использованием технологии NVIDIA CUDA версии 4. В первой части последовательно излагаются основы программной модели CUDA применительно к языкам C и Fortran, сведения о типах памяти GPU и методы эффективного использования разделяемой памяти на примере некоторых вычислительных алгоритмов. Во второй части дан обзор прикладных математических библиотек и языковых надстроек на основе CUDA. Специальные разделы книги посвящены элементам профессиональной разработки – средствам анализа, отладки и диагностики. Рассмотрены методы управления несколькими GPU на рабочих станциях и распределенных кластерных системах. Заключительная часть содержит несколько статей о применении CUDA в задачах математического моделирования гидродинамических процессов и компьютерной графике. Книга предназначена для разработчиков и исследователей, применяющих параллельные вычисления.

Bibtex:

@book{boreskov12parallelnye,
    author = {Боресков, А. В. and Харламов, А. А. and Марковский, Н. Д. and Микушин, Д. Н. and Мортиков, Е. В. and Мыльцев, А. А. and Сахарных, Н. А. and Фролов, В. А.},
    isbn = {978-5-211 06340-2},
    language = {russian},
    numpages = {336},
    publisher = {Москва: МГУ Москва: МГУ},
    title = {Параллельные вычисления на GPU. Архитектура и программная модель CUDA},
    year = {2012}
}

Abstract:

В данной работе рассмотрен алгоритм, основанный на аппроксимации точного расчета глобального освещения. Алгоритм работает в пространстве экрана и основан на идее переиспользования информации о вторичной освещенности в соседних пикселях изображения. Программная реализация использует трассировку лучей в поле расстояний до поверхности и позволяет рассчитывать глобальное освещение интерактивно на современных графических процессорах (2-5 FPS, GTX 570).

Bibtex:

@inproceedings{gruzdev12uskorenie,
    author = {Груздев, А. А. and Фролов, В. А. and Игнатенко, А. В.},
    booktitle = {Труды 22-й Международной конференции по компьютерной графике и зрению ГрафиКон'2012},
    language = {russian},
    pages = {269--271},
    publisher = {Московский государственный университет им. М.В.Ломоносова Москва},
    series = {GraphiCon},
    title = {Ускорение расчёта вторичного освещения с помощью фильтрации в пространстве экрана и уточнения на основе информации о близлежащей геометрии},
    year = {2012}
}

Bibtex:

@inproceedings{petrjajkin12algoritm,
    author = {Петряйкин, А. and Ахадов, Т. and Сенюкова, О. and Крылов, А.},
    booktitle = {Труды международного симпозиума "Нейроимиджинг и магнитоэнцефалография: фундаментальные исследования и клиническая практика"},
    language = {russian},
    pages = {92--94},
    title = {Алгоритм идентификации очагов диффузно-аксонального повреждения у больных с черепно-мозговой травмой},
    year = {2012}
}

Bibtex:

@article{senjukova12vydelenie,
    author = {Сенюкова, О. В. and Галанин, В. Е.},
    issn = {0236-235X},
    journal = {Программные продукты и системы},
    language = {russian},
    number = {1},
    pages = {52--55},
    title = {Выделение областей интереса на основе классификации изолиний},
    year = {2012}
}

Bibtex:

@inproceedings{erofeev11udalenie,
    author = {Ерофеев, М. В.},
    booktitle = {XVIII Международная конференция студентов, аспирантов и молодых ученых "Ломоносов-2011", ISBN 978-5-317-03634-8},
    language = {russian},
    pages = {62--63},
    publisher = {Москва},
    title = {Удаление анимированных и полупрозрачных логотипов из видеоряда},
    year = {2011}
}

Bibtex:

@inproceedings{zeynalov11infrared,
    author = {Zeynalov, R. and Yakubenko, A. and Konushin, A.},
    booktitle = {Proceedings of  8th Open German-Russian Workshop "Pattern Recognition and Image Understanding"},
    language = {english},
    pages = {115--119},
    publisher = {Lobachevsky State University of Nizhny Novgorod},
    series = {OGRW-8-2011},
    title = {Infrared marker matching for object tracking in stereo setup},
    year = {2011}
}

Bibtex:

@inproceedings{kononov11people,
    author = {Kononov, Vladimir and Konushin, Vadim and Konushin, Anton},
    booktitle = {Pattern Recognition},
    doi = {10.1007/978-3-642-23123-0\_17},
    language = {english},
    pages = {163--172},
    publisher = {Springer Berlin Heidelberg},
    series = {Lecture Notes in Computer Science},
    title = {People Tracking Algorithm for Human Height Mounted Cameras},
    volume = {6835},
    year = {2011}
}

Bibtex:

@article{yakubenko11reconstruction,
    address = {Road Town, United Kingdom},
    author = {Yakubenko, A. A. and Kononov, V. A. and Mizin, I. S. and Konushin, V. S. and Konushin, A. S.},
    doi = {10.1134/S0361768811050057},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {5},
    pages = {260--269},
    publisher = {Road Town, United Kingdom},
    title = {Reconstruction of Structure and Texture of City Building Facades},
    volume = {37},
    year = {2011}
}

Bibtex:

@inproceedings{shalnov11algoritm,
    author = {Шальнов, Евгений and Кононов, Владимир and Конушин, Вадим},
    booktitle = {Труды 21-й Международной конференции по компьютерной графике и зрению ГрафиКон'2011},
    language = {russian},
    pages = {227--230},
    publisher = {МАКС Пресс Москва},
    series = {GraphiCon},
    title = {Алгоритм вычитания фона, основанный на поблочных классификаторах},
    year = {2011}
}

Bibtex:

@article{konushin11algoritm,
    author = {КОНУШИН, В. С. and КОНОНОВ, В. А. and КОНУШИН, А. С.},
    issn = {0132-3474},
    journal = {Программирование},
    language = {russian},
    number = {5},
    pages = {27--38},
    title = {Алгоритм слежения за людьми для построения интеллектуального интерфейса человек-компьютер на основе фильтра частиц},
    year = {2011}
}

Bibtex:

@inproceedings{zejnalov11otslezhivanie,
    author = {Зейналов, Р. and Якубенко, А. and Конушин, А.},
    booktitle = {Труды 21-й Международной конференции по компьютерной графике и зрению ГрафиКон'2011},
    language = {russian},
    pages = {263--266},
    publisher = {МАКС Пресс Москва},
    series = {GraphiCon},
    title = {Отслеживание объектов с использование с инфракрасных маркеров},
    year = {2011}
}

Abstract:

Adding a visual logo to a video sequence is a popular method of identifying the owner of that sequence. In this paper we propose afully automatic method for removing opaque, semitransparent and animated logotypes from video sequences.

Bibtex:

@inproceedings{mikhail11automatic,
    author = {Mikhail, E. and Vatolin, D.},
    booktitle = {Proceedings of the 21st International Conference on Computer Graphics and Vision GraphiCon'2011},
    language = {english},
    pages = {204--207},
    publisher = {Moscow, Russia},
    series = {GraphiCon},
    title = {Automatic logo removal for semitransparent and animated logos},
    year = {2011}
}

Abstract:

In this paper, we propose a method of filtering depth maps that are automatically generated from video sequences using optical flow, 3D reconstruction and scene analysis methods. To attain better quality, information from both the source video and depth map is used. The proposed algorithm uses motion estimation to take into account temporal information, but the algorithm’s structure permits use of optical flow to improve quality, but at the expense of greater computation time. The method can be applied as a preprocessing stage for enhancement of multi-view or stereo video. Joint temporal and spatial processing can yield further improvements in quality. A comparison of the results with test ground-truth sequences using the BI-PSNR metric is presented.

Bibtex:

@inproceedings{matyunin11temporal,
    author = {Matyunin, Sergey and Vatolin, Dmitriy and Smirnov, Maxim},
    booktitle = {9th WSCG International Conference on Computer Graphics, Visualization and Computer Vision},
    language = {english},
    pages = {5--8},
    series = {WSCG},
    title = {Fast temporal filtering of depth maps},
    year = {2011}
}

Bibtex:

@inproceedings{voronov11novel,
    author = {Voronov, Alexander and Vatolin, Dmitriy and Smirnov, Maxim},
    booktitle = {9th WSCG International Conference on Computer Graphics, Visualization and Computer Vision},
    language = {english},
    pages = {25--30},
    title = {Novel trilateral approach for depth map spatial filtering},
    year = {2011}
}

Abstract:

In this paper we present our approach to real-time filtering of depth maps taken using projected-pattern-based depth cameras and to restoration of scene backgrounds for images taken using aligned RGB and depth cameras. An original depth map contains a numerous occlusions, and stereo-from-depth map video generation leaves many uncovered areas. To solve this problem we propose an adaptive occlusion-filling algorithm for depth map processing and for restoration of scene backgrounds using depth map. Our goal is to accurately fill occlusions while maintaining real-time processing speed using common workstations.

Bibtex:

@inproceedings{berdnikov11realtime,
    author = {Berdnikov, Y. and Vatolin, D.},
    booktitle = {21st International Conference on Computer Graphics and Vision, GraphiCon'2011 - Conference Proceedings},
    language = {english},
    pages = {200--203},
    publisher = {Moscow, Russia},
    series = {GraphiCon},
    title = {Real-time depth map occlusion filling and scene background restoration for projected-pattern-based depth cameras},
    year = {2011}
}

Abstract:

This paper presents a novel approach for depth map estimation from a single image using information about edge blur. The blur amount at the edge is calculated from the gradient magnitude ratio of the input and re-blurred images. The final depth map can be obtained by propagating estimated information from the edges to the entire image using cross-bilateral filtering. Experimental results for real images and video sequences demonstrate the effectiveness of this method in providing a plausible depth map for 2D-to-3D conversion that generates comfortable stereo viewing.

Bibtex:

@inproceedings{akimov11singleimage,
    author = {Akimov, Dmitry and Vatolin, Dmitriy and Smirnov, Maxim},
    booktitle = {Proceedings of the 21st International Conference on Computer Graphics and Vision GraphiCon'2011},
    language = {english},
    pages = {12--15},
    publisher = {Moscow, Russia},
    series = {GraphiCon},
    title = {Single-Image Depth Map Estimation Using Blur Information},
    year = {2011}
}

Bibtex:

@inproceedings{matyunin11temporalpostprocessing,
    author = {Matyunin, Sergey and Vatolin, Dmitry and Smirnov, Michail},
    booktitle = {International Conference on Imaging Theory and Applications},
    language = {english},
    pages = {33--38},
    title = {Temporal Post-processing Method for Automatically Generated Depth Maps},
    year = {2011}
}

Abstract:

We propose a method of filtering depth maps provided by Kinect depth camera. Filter uses output of the conventional Kinect camera along with the depth sensor to improve the temporal stability of the depth map and fill occlusion areas. To filter input depth map, the algorithm uses the information about motion and color of objects from the video. The proposed method can be applied as a preprocessing stage before using Kinect output data.

Bibtex:

@inproceedings{matyunin11temporalfiltering,
    author = {Matyunin, Sergey and Vatolin, Dmitry and Berdnikov, Yury and Smirnov, Maxim},
    booktitle = {3DTV Conference: The True Vision-Capture, Transmission and Display of 3D Video (3DTV-CON), 2011},
    doi = {10.1109/3DTV.2011.5877202},
    language = {english},
    pages = {1--4},
    title = {Temporal filtering for depth maps generated by Kinect depth camera},
    year = {2011}
}

Bibtex:

@inproceedings{zachesov11avtomaticheskij,
    author = {Зачесов, А. А. and Колиниченко, М. Б. and Ватолин, Д. С.},
    booktitle = {Научно-техническая конференция "Теоретические и технические вопросы записи, воспроизведения и восприятия стереоизображений"},
    language = {russian},
    pages = {14--16},
    title = {Автоматический алгоритм построения многоракурсного видео по стереопаре},
    year = {2011}
}

Bibtex:

@inproceedings{vatolin11vysokoeffektivnoe,
    author = {Ватолин, Д. С. and Матюнин, С. Б.},
    booktitle = {Научно-техническая конференция "Теоретические и технические вопросы записи, воспроизведения и восприятия стереоизображений"},
    language = {russian},
    pages = {17--19},
    title = {Высокоэффективное сжатие карт глубины с использованием информации 2D видеопотока},
    year = {2011}
}

Bibtex:

@inproceedings{sumin11opredelenie,
    author = {Сумин, Д. А. and Федоров, А. А. and Ватолин, Д. С.},
    booktitle = {Научно-техническая конференция "Теоретические и технические вопросы записи, воспроизведения и восприятия стереоизображений"},
    language = {russian},
    pages = {6--8},
    title = {Определение характеристик устройств отображения 3D-видео},
    year = {2011}
}

Bibtex:

@article{frolov11biased,
    address = {Road Town, United Kingdom},
    author = {Frolov, V. A. and Kharlamov, A. A. and Ignatenko, A.},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {5},
    pages = {252--259},
    publisher = {Road Town, United Kingdom},
    title = {Biased solution of integral illumination equation via irradiance caching and path tracing on GPUs},
    volume = {37},
    year = {2011}
}

Abstract:

В данной работе представлен подход к синтезу фотореалистичных изображений на графических процессорах (GPU). В его основе лежит сочетание алгоритмов кэширования освещѐнности с когерентной адаптивной трассировкой путей.

Bibtex:

@article{frolov11smeschennoe,
    author = {Фролов, В. А. and Харламов, А. А. and Игнатенко, А. В.},
    doi = {10.1134/S0361768811050021},
    issn = {0132-3474},
    journal = {Программирование},
    language = {russian},
    number = {5},
    pages = {47--60},
    title = {Смещённое решение интегрального уравнения светопереноса на графических процессорах при помощи трассировки путей и кэша освещенности},
    volume = {37},
    year = {2011}
}

Bibtex:

@article{senyukova11automated,
    address = {Road Town, United Kingdom},
    author = {Senyukova, O. V. and Lukin, A. S. and Vetrov, D. P.},
    doi = {10.1134/S0361768811050045},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {5},
    pages = {245--251},
    publisher = {Road Town, United Kingdom},
    title = {Automated Atlas-Based Segmentation of NISSL-Stained Mouse Brain Sections Using Supervised Learning},
    volume = {37},
    year = {2011}
}

Bibtex:

@inproceedings{senyukova11diagnostics,
    author = {Senyukova, O. V. and Gavrishchaka, V. V.},
    booktitle = {Proceedings of International Conference on Computer and Computational Intelligence},
    doi = {10.1115/1.859926.paper4},
    language = {english},
    pages = {19--26},
    title = {Diagnostics of complex and rare abnormalities using ensemble decomposition learning},
    year = {2011}
}

Bibtex:

@inproceedings{senyukova11diffuse,
    author = {Senyukova, O. V. and Galanine, V. E. and Krylov, A. S. and Petraikin, A. V. and Akhadov, T. A. and Sidorin, S. V.},
    booktitle = {Proceedings of the 21st International Conference on Computer Graphics and Vision GraphiCon'2011},
    language = {english},
    pages = {84--87},
    publisher = {Moscow, Russia},
    series = {GraphiCon},
    title = {Diffuse Axonal Injury Lesion Segmentation Using Contouring Algorithm},
    year = {2011}
}

Bibtex:

@inproceedings{senyukova11ensemble,
    author = {Senyukova, O. and Gavrishchaka, V.},
    booktitle = {Proceedings of IASTED International Conference on Computational Intelligence and Bioinformatics},
    doi = {10.2316/P.2011.753-025},
    language = {english},
    pages = {69--73},
    title = {Ensemble Decomposition Learning for Optimal Utilization of Implicitly Encoded Knowledge in Biomedical Applications},
    year = {2011}
}

Bibtex:

@inproceedings{gavrishchaka11multiobjective,
    author = {Gavrishchaka, V. and Senyukova, O. and Koepke, M. and Kryuchkova, A.},
    booktitle = {Proceedings of International Conference on Computer and Computational Intelligence},
    doi = {10.1115/1.859926.paper16},
    language = {english},
    pages = {95--106},
    title = {Multi-objective physiological indicators based on complementary complexity measures: application to early diagnostics and prediction of acute events},
    year = {2011}
}

Bibtex:

@inproceedings{gavrishchaka11physiological,
    author = {Gavrishchaka, V. and Senyukova, O. and Ulanova, O. and Monin, A.},
    booktitle = {Proceedings of VII International scientific and practical conference for memory of P.Roudik},
    language = {english},
    pages = {5--7},
    title = {Physiological meta-indicators for professional sports applications: express diagnostics, overtraining detection, and quantification of individual zones of optimal functioning},
    year = {2011}
}

Bibtex:

@inproceedings{gavrishchaka11robust,
    author = {Gavrishchaka, V. and Senyukova, O.},
    booktitle = {2011 INTERNATIONAL SYMPOSIUM ON COMPUTATIONAL MODELS FOR LIFE SCIENCES (CMLS-11)},
    doi = {10.1063/1.3596645},
    language = {english},
    pages = {215--224},
    series = {AIP Conference Proceedings},
    title = {Robust algorithmic detection of the developed cardiac pathologies and emerging or transient abnormalities from short periods of RR data},
    volume = {1371},
    year = {2011}
}

Bibtex:

@article{senjukova11avtomaticheskaja,
    author = {Сенюкова, О. В. and Лукин, А. С. and Ветров, Д. П.},
    issn = {0132-3474},
    journal = {Программирование},
    language = {russian},
    number = {5},
    pages = {39--48},
    title = {Автоматическая сегментация срезов мозга мыши, окрашенных по NISLL, основанная на обучении с учителем по разметке из атласа},
    year = {2011}
}

Bibtex:

@inproceedings{yakubenko10automatic,
    author = {Yakubenko, A. and Mizin, I. and Konushin, A.},
    booktitle = {Proceedings of the 20th International Conference on Computer Graphics and Vision GraphiCon'2010},
    language = {english},
    pages = {100--106},
    publisher = {St. Petersburg, Russia},
    series = {GraphiCon},
    title = {Automatic Extraction of Regular Grids from Rectified Facade Image},
    year = {2010}
}

Bibtex:

@inproceedings{konushin10improvement,
    author = {Konushin, V. and Konushin, A.},
    booktitle = {Proceedings of the 20th International Conference on Computer Graphics and Vision GraphiCon'2010},
    language = {english},
    pages = {96--99},
    publisher = {St. Petersburg, Russia},
    series = {GraphiCon},
    title = {Improvement of background subtraction by mask constraints},
    year = {2010}
}

Bibtex:

@article{konushin10algoritm,
    address = {М.},
    author = {Конушин, В. and Кривовязь, Г. and Конушин, А.},
    issn = {1992-2264; 2310-9912},
    journal = {Информатика и ее применения},
    language = {russian},
    number = {1},
    pages = {74--78},
    publisher = {М.},
    title = {Алгоритм распознавания людей в видеопоследовательности на основе случайных патчей},
    volume = {4},
    year = {2010}
}

Bibtex:

@inproceedings{sindeev10otslezhivanie,
    author = {Синдеев, М. and Конушин, А.},
    booktitle = {Труды 20-й Международной конференции по компьютерной графике и зрению ГрафиКон'2010},
    language = {russian},
    pages = {214--217},
    publisher = {ИТМО Санкт-Петербург},
    series = {GraphiCon},
    title = {Отслеживание контура лабораторной мыши в видеопоследовательности на основе метода ветвей и границ},
    year = {2010}
}

Bibtex:

@article{jakubenko10poisk,
    author = {Якубенко, А. А. and Мизин, И. С. and Конушин, А. С.},
    issn = {0236-235X},
    journal = {Программные продукты и системы},
    language = {russian},
    number = {4},
    pages = {162--167},
    title = {Поиск регулярных решеток на текстуре фасадов зданий},
    year = {2010}
}

Bibtex:

@inproceedings{chigorin10postroenie,
    author = {Чигорин, А. and Конушин, А.},
    booktitle = {Труды 20-й Международной конференции по компьютерной графике и зрению ГрафиКон'2010},
    language = {russian},
    pages = {348--350},
    publisher = {ИТМО Санкт-Петербург},
    series = {GraphiCon},
    title = {Построение трехмерной модели порового пространства горных пород},
    year = {2010}
}

Bibtex:

@inproceedings{vatolin10temporalnyj,
    author = {Ватолин, Д. В. and Матюнин, С. Б. and Смирнов, М. А.},
    booktitle = {II Международная научно-практическая конференция "Молодежь. Наука. Инновации"},
    language = {russian},
    pages = {229--233},
    title = {Tемпоральный метод фильтрации карт глубины},
    year = {2010}
}

Bibtex:

@inproceedings{vatolin10trilateralnaja,
    author = {Ватолин, Д. С. and Воронов, А. А. and Смирнов, М. А.},
    booktitle = {II Международная научно-практическая конференция "Молодежь. Наука. Инновации"},
    language = {russian},
    pages = {224--229},
    title = {Трилатеральная постфильтрация карт глубины},
    year = {2010}
}

Bibtex:

@inproceedings{frolov10fotorealistichnaja,
    author = {Фролов, В. А. and Игнатенко, А. В.},
    booktitle = {Новые информационные технологии в автоматизированных системах. Материалы 13-ого научно-практического семинара. M.: МГИЭМ},
    language = {russian},
    pages = {149--150},
    title = {Фотореалистичная визуализация с помощью трассировки путей на графических процессорах},
    year = {2010}
}

Abstract:

Сенюкова О, Конушин А, Ветров Д, Анохин К. Поверхностная визуализация данных об экспресии генов в коре головного мозга и гиппокампе мозга мыши. Труды 12-й всероссийской научно-технической конференции "Нейроинформатика-2010", 1, 249-257, 2010.

Bibtex:

@inproceedings{senyukova10automated,
    author = {Senyukova, O. and Lukin, A. and Vetrov, D.},
    booktitle = {Proceedings of the 20th International Conference on Computer Graphics and Vision GraphiCon'2010},
    language = {english},
    pages = {92--96},
    publisher = {St. Petersburg, Russia},
    series = {GraphiCon},
    title = {Automated Atlas-Based Segmentation of Nissl-Stained Mouse Brain Slices},
    year = {2010}
}

Bibtex:

@inproceedings{senjukova10poverhnostnaja,
    author = {Сенюкова, О. and Конушин, А. and Ветров, Д. and Анохин, К.},
    booktitle = {Труды 12-й всероссийской научно-технической конференции "Нейроинформатика"},
    language = {russian},
    pages = {249--257},
    title = {Поверхностная визуализация данных об экспресии генов в коре головного мозга и гиппокампе мозга мыши},
    volume = {1},
    year = {2010}
}

Abstract:

The paper describes a method for fully automatic 3D-reconstruction of mouse brain voxel model from a sequence of coronal 2D slices for statistical analysis of gene expression. Two images of each brain slice with different stains are used. The first stain highlights the histology of brain, which is used for slice matching. The second stain highlights the level of gene expression. The algorithm proceeds as follows. First, images are preprocessed to suppress image noise and equalize image brightness. Second we estimate the level of gene expression in each slice using the second stain. Then we construct 3D-model of the brain using the first stain. To do this all images are aligned via rigid-body transformations. After alignment neighboring slices are matched by estimation of non-linear deformations. As the distance between slices is significantly larger then image resolution we add intermediate virtual slices using morphing algorithm. Gene expression level is interpolated in identical way. The obtained 3D-model with the information about gene expression can be used for gene expression analysis via Statistical Parametric Mapping (SPM) package. The proposed method for 3D-reconstruction has been tested on images from Allen Brain Atlas, which is available in electronic form.

Bibtex:

@inproceedings{osokin09adaptation,
    author = {Osokin, A. and Lebedev, A. and Vetrov, D. and Galatenko, V. and Kropotov, D. and Nedzved, A. and Konushin, A. and Anokhin, K.},
    booktitle = {Proceedings of the 19th International Conference on Computer Graphics and Vision GraphiCon'2009},
    language = {english},
    pages = {42--48},
    publisher = {MAKS Press Moscow, Russia},
    series = {GraphiCon},
    title = {Adaptation of mouse brain gene expression data for further Statistical Parametrical Mapping analysis},
    year = {2009}
}

Bibtex:

@inproceedings{barinova09efficient,
    author = {Barinova, O. and Shapovalov, R. and Sudakov, S. and Velizhev, A. and Konushin, A.},
    booktitle = {Proceedings of 3D City Models, Road Databases and Traffic Monitoring (CMRT-2009)},
    language = {english},
    pages = {1--6},
    series = {IAPRS},
    title = {Efficient road mapping via interactive image segmentation},
    year = {2009}
}

Bibtex:

@article{lomakinarumyantseva09video,
    address = {Road Town, United Kingdom},
    author = {Lomakina-Rumyantseva, E. and Voronin, P. and Kropotov, D. and Vetrov, D. and Konushin, A.},
    issn = {1555-6212; 1054-6618},
    journal = {Pattern Recognition and Image Analysis: Advances in Mathematical Theory and Applications},
    language = {english},
    number = {4},
    pages = {616--622},
    publisher = {Road Town, United Kingdom},
    title = {Video tracking and behaviour segmentation of laboratory rodents},
    volume = {19},
    year = {2009}
}

Bibtex:

@inproceedings{velizhev09avtomaticheskaja,
    author = {Велижев, А. and Шаповалов, Р. and Потапов, Д. and Третьяк, Е. and Конушин, А.},
    booktitle = {Труды 19-й Международной конференции по компьютерной графике и зрению ГрафиКон'2009},
    language = {russian},
    pages = {241--245},
    publisher = {Москва},
    series = {GraphiCon},
    title = {Автоматическая сегментация облаков точек на основе элементов поверхности},
    year = {2009}
}

Bibtex:

@inproceedings{konushin09algoritm,
    author = {Конушин, В. and Кривовязь, Г. and Конушин, А.},
    booktitle = {Труды 19-й Международной конференции по компьютерной графике и зрению ГрафиКон'2009},
    language = {russian},
    pages = {219--222},
    publisher = {Москва},
    series = {GraphiCon},
    title = {Алгоритм распознавания людей в видеопоследовательности на основе случайных патчей},
    year = {2009}
}

Bibtex:

@inproceedings{kononov09ochischenie,
    author = {Кононов, В. and Конушин, В. and Якубенко, А. and Конушин, А.},
    booktitle = {Труды 19-й Международной конференции по компьютерной графике и зрению ГрафиКон'2009},
    language = {russian},
    pages = {283--286},
    publisher = {Москва},
    series = {GraphiCon},
    title = {Очищение текстур фасадов зданий с использованием их структуры},
    year = {2009}
}

Abstract:

An algorithm for video sequence frame rate double up-conversion is described. The algorithm is based on bidirectional motion compensation. The algorithm has no smoothing in the time domain implying the absence of the image sharpness oscillations in the transformed video sequence. A special post-processing step with adaptively controlled degree of smoothing makes it possible to considerably decrease the “blocking” artifact while retaining the maximum number of image details. The absence of complicated mathematical computations allows realtime hardware implementation of the algorithm and real-time video processing.

Bibtex:

@article{vatolin09double,
    address = {Road Town, United Kingdom},
    author = {Vatolin, D. S. and Grishin, S. V.},
    doi = {10.1134/S036176880906005X},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {6},
    pages = {351--364},
    publisher = {Road Town, United Kingdom},
    title = {Double up-conversion of video frame rate based on bidirectional motion compensation},
    volume = {35},
    year = {2009}
}

Abstract:

The problem of image interpolation is one of the most thoroughly developed in the area of multimedia processing. Linear interpolation methods suffer from aliasing and Hibbs effect, and they fail to provide artifacts free enlargement of edges. This has led to a plethora of edge-oriented upscaling techniques aimed at providing superior edge processing quality. However, Edge-Directed Interpolation (EDI) of textured areas results in strong artifacts, making the discrimination between edges and texture an important task as well.

Bibtex:

@inproceedings{simonyan09edgedirected,
    author = {Simonyan, Karen and Vatolin, Dmitriy},
    booktitle = {Proceedings of the British Machine Vision Conference},
    doi = {10.5244/C.23.50},
    language = {english},
    pages = {1521--1527},
    title = {Edge-Directed Interpolation in a Bayesian Framework},
    volume = {10},
    year = {2009}
}

Abstract:

In this paper we propose a novel approach to video codec evaluation, comparison and testing based on preparation of specific video sequences followed by efficiency evaluation. Several strategies for applying natural video sequence modification as well as fully artificial video clip creation are suggested and studied. Experimental results with measurements and scoring summary for several MPEG-2, MPEG-4 ASP and H.264 codecs are presented and discussed.

Bibtex:

@inproceedings{parshin09novel,
    author = {Parshin, Alexander and Vatolin, Dmitry and Brailovskiy, Ilya and Corriveau, Philip},
    booktitle = {Proceedings of the Fourth International Workshop on Video Processing and Quality Metrics for Consumer Electronics},
    language = {english},
    title = {Novel Approach for Video Codecs Scoring using Specially Prepared Video Sequences},
    year = {2009}
}

Bibtex:

@inproceedings{grishin09algoritm,
    author = {Гришин, СВ and Симонян, КА and Ватолин, ДС},
    booktitle = {Материалы двенадцатого научно-практическо го семинара "Новые информационные технологии в автоматизирован ных системах"},
    language = {russian},
    pages = {19--29},
    publisher = {Москва},
    title = {Алгоритм вычисления параметров наложений для задачи преобразования частоты кадров цифровых видеосигналов},
    year = {2009}
}

Bibtex:

@article{grishin09bystryj,
    author = {Гришин, СВ and Ватолин, ДС},
    issn = {0236-235X},
    journal = {Программные продукты и системы},
    language = {russian},
    pages = {159--63},
    title = {Быстрый алгоритм преобразования частоты кадров},
    volume = {2},
    year = {2009}
}

Bibtex:

@article{vatolin09dvukratnoe,
    author = {Ватолин, ДС and Гришин, СВ},
    issn = {0132-3474},
    journal = {Программирование},
    language = {russian},
    number = {6},
    pages = {67--80},
    title = {Двукратное увеличение частоты кадров видео на основе двунаправленной компенсации движения},
    volume = {35},
    year = {2009}
}

Abstract:

Данная работа посвящена ускорению алгоритма фотонных карт с использованием графических процессоров (GPU). Фотонные карты - один из наиболее универсальных и реалистичных методов синтеза изображений, но и один из самых требовательных к вычислительным ресурсам. Мы представляем модифицированный алгоритм, работающий на CUDA-совместимых GPU. В работе также предложен подход организации трассировки лучей, позволяющий в высокой степени утилизировать ресурсы GPU. Наша реализация обратной трассировки лучей такая же быстрая, как и в передовых работах по интерактивной трассировке лучей на GPU.

Bibtex:

@inproceedings{frolov09interaktivnaja,
    author = {Фролов, В. А. and Игнатенко, А. В.},
    booktitle = {Труды 19-й Международной конференции по компьютерной графике и зрению ГрафиКон'2009},
    language = {russian},
    pages = {255--262},
    publisher = {Москва},
    series = {GraphiCon},
    title = {Интерактивная трассировка лучей и фотонные карты на GPU},
    year = {2009}
}

Bibtex:

@inproceedings{senyukova09visualization,
    author = {Senyukova, O.},
    booktitle = {Proceedings of the 19th International Conference on Computer Graphics and Vision GraphiCon'2009},
    language = {english},
    pages = {372--375},
    publisher = {MAKS Press Moscow, Russia},
    series = {GraphiCon},
    title = {Visualization of gene expression data on the surface of mouse brain cerebral cortex and hippocampus},
    year = {2009}
}

Bibtex:

@inproceedings{senjukova09vizualizatsija,
    author = {Сенюкова, О. and Конушин, А. and Ветров, Д. and Анохин, К.},
    booktitle = {Материалы Международной конференции "Современные проблемы математики, механики и их приложений"(Москва, 30-2 апреля 2009)},
    language = {russian},
    pages = {330--331},
    publisher = {изд.-во Университетская книга Москва},
    title = {Визуализация активности нейронов на поверхности мозга мыши},
    year = {2009}
}

Bibtex:

@inproceedings{konushin08automated,
    author = {Konushin, Anton and Kropotov, Dmitry and Vetrov, Dmitry and Cherepov, Anton and V, Anokhin Konstantin},
    booktitle = {Proceedings of Measuring Behavior 2008, 6th International Conference on Methods and Techniques in Behavioral Research},
    language = {english},
    pages = {26--29},
    publisher = {Publisher: Wageningen: Noldus Information Technology Maastricht, The Netherlands},
    title = {Automated distinguishing of mouse behavior in new environment and under amphetamine using decision trees},
    year = {2008}
}

Bibtex:

@inproceedings{konushin08automateddistinguishing,
    author = {Konushin, A. and Lomakina-Rumyantseva, E. and Kropotov, D. and Vetrov, D. and Cherepov, A. and Anokhin, K.},
    booktitle = {Proc. of Measuring Behavior - 2008},
    language = {english},
    pages = {304--305},
    publisher = {Maastriht, Holland},
    title = {Automated distinguishing of mouse behavior in new environment and under amphetemine using decision trees},
    year = {2008}
}

Bibtex:

@inproceedings{vetrov08automatic,
    author = {Vetrov, D. and Kropotov, D. and Konushin, A. and Lomakina-Rumyantseva, E. and Zarayskaya, I. and Anokhin, K.},
    booktitle = {Proc. of Measuring Behavior - 2008},
    language = {english},
    pages = {241--242},
    publisher = {Maastriht, Holland},
    title = {Automatic segmentation of mouse behavior using hidden Markov models},
    year = {2008}
}

Bibtex:

@inproceedings{konushin08behavior,
    author = {Konushin, A. and Vetrov, D. and Voronin, P. and Sindeev, M. and Lomakina-Rumyantseva, E. and Kropotov, D. and Anokhin, K. and Zaraiskaya, I.},
    booktitle = {GraphiCon 2008 - International Conference on Computer Graphics and Vision, Proceedings},
    language = {english},
    title = {BEHAVIOR VIDEO TRACKING SYSTEM WITH AUTOMATIC SEGMENTATION INTO BEHAVIORIAL ACTS},
    year = {2008}
}

Abstract:

Video tracking systems are widely used for behavior analysis of laboratory animals. They perform tracking of 2-3 typical points of animals. For further analysis it is required to identify separate behavioral acts. Currently this can be done only manually by a specialist in animal behavior. Existing automatic methods are able to mark the periods of immobility and active motion. They have relatively low accuracy and require careful parameter setting. In the paper we suggest video tracking system which models animal behavior by using hidden Markov models. The system shows promising results in analyzing mouse behavior in the new environment.

Bibtex:

@inproceedings{konushin08behaviorvideo,
    author = {Konushin, A. and Vetrov, D. and Kropotov, D. and Voronin, P. and Sindeev, M. and Lomakina-Rumyantseva, E. and Anokhin, K. and Zaraiskaya, I.},
    booktitle = {Proceedings of the 18th International Conference on Computer Graphics and Vision GraphiCon'2008},
    language = {english},
    publisher = {Moscow, Russia},
    series = {GraphiCon},
    title = {Behavior video tracking system with automatic segmentation into behaviorial acts},
    year = {2008}
}

Bibtex:

@inproceedings{barinova08automatic,
    author = {Barinova, Olga and Konushin, Vadim and Yakubenko, Anton and Lee, KeeChang and Lim, Hwasup and Konushin, Anton},
    booktitle = {Computer Vision – ECCV 2008},
    doi = {10.1007/978-3-540-88688-4\_8},
    language = {english},
    pages = {100--113},
    publisher = {Springer Berlin Heidelberg},
    series = {Lecture Notes in Computer Science},
    title = {Fast Automatic Single-View 3-d Reconstruction of Urban Scenes},
    volume = {5303},
    year = {2008}
}

Bibtex:

@inproceedings{sudakov08semantic,
    author = {Sudakov, S. and Barinova, O. and Velizhev, A. and Konushin, A.},
    booktitle = {Proceedings of Pattern recognition and image analysis-2008},
    language = {english},
    pages = {182--185},
    title = {Semantic segmentation of road images based on cascade classifiers},
    volume = {2},
    year = {2008}
}

Bibtex:

@inproceedings{voronin08video,
    author = {Voronin, P. and Konushin, A.},
    booktitle = {Proceedings of Pattern recognition and image analysis-2008},
    language = {english},
    pages = {299--303},
    title = {Video tracking of laboratory rodents with active shape models},
    volume = {2},
    year = {2008}
}

Bibtex:

@inproceedings{sudakov08algoritmy,
    author = {Судаков, С. and Семашко, А. and Баринова, О. and Конушин, А. and Киншаков, В. and Крылов, А.},
    booktitle = {Труды 18-й Международной конференции по компьютерной графике и зрению ГрафиКон'2008},
    language = {russian},
    pages = {206--212},
    publisher = {Москва},
    series = {GraphiCon},
    title = {Алгоритмы детектирования разметки и дефектов дорожного покрытия},
    year = {2008}
}

Bibtex:

@inproceedings{barinova08bystraja,
    author = {Баринова, О. and Конушин, В. and Соболев, А. and Кузьмишкина, А. and Якубенко, А. and Лим, Х. and Конушин, А.},
    booktitle = {Труды 18-й Международной конференции по компьютерной графике и зрению ГрафиКон'2008},
    language = {russian},
    pages = {234--241},
    publisher = {Москва},
    series = {GraphiCon},
    title = {Быстрая автоматическая трехмерная реконструкция городских сцен по одному изображению},
    year = {2008}
}

Bibtex:

@inproceedings{barinova08bystryj,
    author = {Баринова, Ольга and Конушин, Вадим and Якубенко, Антон and Конушин, Антон},
    booktitle = {Интеллектуализация обработки информации: Тезисы докладов Международной научной конференции},
    language = {russian},
    pages = {22--24},
    publisher = {Крымский научный центр НАН Украины Симферополь},
    title = {Быстрый метод семантической сегментации изображений для автоматической трехмерной реконструкции городских сцен по одной фотографии},
    year = {2008}
}

Bibtex:

@inproceedings{barinova08bystryjmetod,
    author = {Баринова, Ольга and Конушин, Вадим and Якубенко, Антон and Конушин, Антон},
    booktitle = {Интеллектуализация обработки информации: Тезисы докладов Международной научной конференции},
    language = {russian},
    pages = {22--24},
    publisher = {Крымский научный центр НАН Украины Симферополь},
    title = {Быстрый метод семантической сегментации изображений для автоматической трехмерной реконструкции городских сцен по одной фотографии},
    year = {2008}
}

Bibtex:

@inproceedings{kuzmishkina08vybor,
    author = {Кузьмишкина, А. and Баринова, О. and Конушин, А.},
    booktitle = {Труды 18-й Международной конференции по компьютерной графике и зрению ГрафиКон'2008},
    language = {russian},
    pages = {213--216},
    publisher = {Москва},
    series = {GraphiCon},
    title = {Выбор объектов для обучения в условиях сильной несбалансированности классов},
    year = {2008}
}

Bibtex:

@inproceedings{shapovalov08onlajnbegging,
    author = {Шаповалов, Р. В. and Баринова, О. В. and Велижев, А. Б. and Конушин, А. С.},
    booktitle = {Труды 51ой конференции МФТИ "Современные проблемы фундаментальных и прикладных наук"},
    language = {russian},
    pages = {79--81},
    title = {Онлайн-бэггинг, чувствительный к стоимостям ошибок на разных классах},
    volume = {3},
    year = {2008}
}

Bibtex:

@inproceedings{konushin08sistema,
    author = {Конушин, А. and Ветров, Д. and Воронин, П. and Синдеев, М. and Ломакина-Румянцева, Е.},
    booktitle = {Труды 18-й Международной конференции по компьютерной графике и зрению ГрафиКон'2008},
    language = {russian},
    pages = {199--205},
    publisher = {Москва},
    series = {GraphiCon},
    title = {Система видеонаблюдения за поведением лабораторных животных с автоматической сегментацией на поведенческие акты},
    year = {2008}
}

Abstract:

In this paper we propose a confidence measure for block-based motion vector field. The measure is calculated as an average of two a posteriori estimates which reflect various aspects of MVF accuracy: motion-compensated interframe difference distribution and motion vectors distribution. Experimental results show that the proposed measure outperforms its contemporary counterpart while demanding less information about the motion estimation process.

Bibtex:

@inproceedings{simonyan08confidence,
    author = {Simonyan, K. and Grishin, S. and Vatolin, D.},
    booktitle = {Proceedings of the 18th International Conference on Computer Graphics and Vision GraphiCon'2008},
    language = {english},
    pages = {110--113},
    publisher = {Moscow, Russia},
    series = {GraphiCon},
    title = {Confidence measure for block-based motion vector field},
    year = {2008}
}

Abstract:

In this paper we propose a novel super-resolution algorithm based on motion compensation and edge-directed spatial interpolation succeeded by fusion via pixel classification. Two high-resolution images are constructed, the first by means of motion compensation and the second by means of edge-directed interpolation. The AdaBoost classifier is then used to fuse these images into an high-resolution frame. Experimental results show that the proposed method surpasses well-known resolution enhancement methods while maintaining moderate computational complexity.

Bibtex:

@inproceedings{simonyan08video,
    author = {Simonyan, Karen and Grishin, Sergey and Vatolin, Dmitriy and Popov, Dmitriy},
    booktitle = {Image Processing, 2008. ICIP 2008. 15th IEEE International Conference on},
    doi = {10.1109/ICIP.2008.4711763},
    language = {english},
    pages = {349--352},
    title = {Fast video super-resolution via classification},
    year = {2008}
}

Abstract:

In this paper we present a novel approach to fully automatic pan- tilt-zoom (PTZ) camera calibration. A calibration s ystem based on the proposed method is easy to setup, is easily sca lable and does not require any human participation during operatio n. We focus on calibration of extrinsic parameters while assumi ng that all intrinsic parameters are known a priori . Our calibration technique uses a set of measurements that are each represente d by the correspondence between the Cartesian world coordina tes and the camera’s internal pan - tilt coordinates for a given point. Use of the proposed “direct measurement” approach makes the ca libration process separable, which means that the camera posi tion and rotation can be calculated independently. Although the internal coordinates are easily accessible in most contempor ary cameras, the world coordinates of specific points must be ob tained using a visual marker detection system. The output of the a uto-calibration system for each camera is the extrinsic camera para meters that best fit the input measurements. The proposed metho d is designed for operation in areas under various lighting condi tions or with complex wall topology and it exhibits adequate resu lts in practice.

Bibtex:

@inproceedings{obukhov08fully,
    author = {Obukhov, Anton and Strelnikov, Konstantin and Vatolin, Dmitry},
    booktitle = {Proceedings of the 18th International Conference on Computer Graphics and Vision GraphiCon'2008},
    language = {english},
    pages = {122--127},
    publisher = {Moscow, Russia},
    series = {GraphiCon},
    title = {Fully Automatic PTZ Camera Calibration Method},
    year = {2008}
}

Abstract:

In this paper, we propose a super-resolution algorithm based on image fusion via pixel classification. Two high-resolution images are constructed, the first by means of motion compensation and the second by means of image interpolation. The AdaBoost classifier is then used in the fusion of these images, resulting in an high-resolution frame. Experimental results show that the proposed method outperforms well-known video resolution enhancement methods while maintaining moderate computational complexity.

Bibtex:

@inproceedings{simonyan08videosuperresolution,
    author = {Simonyan, Karen and Grishin, Sergey and Vatolin, Dmitriy and Popov, Dmitriy},
    booktitle = {Proceedings of the 24th Spring Conference on Computer Graphics},
    doi = {10.1145/1921264.1921294},
    language = {english},
    pages = {143--148},
    title = {Video super-resolution using motion compensation and classification-aided fusion},
    year = {2008}
}

Bibtex:

@inproceedings{parshin08metody,
    author = {Паршин, Александр and Ватолин, Дмитрий},
    booktitle = {Proceedings of the 18th International Conference on Computer Graphics and Vision GraphiCon'2008},
    language = {russian},
    pages = {294--301},
    publisher = {Moscow, Russia},
    series = {GraphiCon},
    title = {Методы анализа видекодеков на основе модификации естественных видеопоследовательностей},
    year = {2008}
}

Bibtex:

@inproceedings{grishin08obzor,
    author = {Гришин, СВ and Ватолин, ДС and Лукин, АС and Путилин, СЮ and Стрельников, КН},
    booktitle = {Программные системы и инструменты. Тематический сборник},
    editor = {Королев, Л.Н. and Корухова, Людмила Сергеевна and Костенко, Валерий Алексеевич},
    language = {russian},
    pages = {50--62},
    publisher = {МАКС Пресс Москва},
    series = {Программные системы и инструменты},
    title = {Обзор блочных методов оценки движения в цифровых видео сигналах},
    volume = {9},
    year = {2008}
}

Bibtex:

@inproceedings{senjukova08trehmernoe,
    author = {Сенюкова, О. and Якубенко, А.},
    booktitle = {Труды XV Международной конференции студентов, аспирантов и молодых учёных "Ломоносов"},
    language = {russian},
    pages = {77},
    title = {Трехмерное моделирование тел формы обобщенного цилиндра по изображениям},
    year = {2008}
}

Bibtex:

@inproceedings{vezhnevets07interactive,
    author = {Vezhnevets, V. and Konushin, A. and Ignatenko, A.},
    booktitle = {The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Science},
    language = {english},
    pages = {63--68},
    series = {ISPRS},
    title = {Interactive Image-Based Urban Modelling},
    volume = {36},
    year = {2007}
}

Bibtex:

@inproceedings{kutuzova07sistema,
    author = {Кутузова, В. and Конушин, А.},
    booktitle = {Материалы XIV Международной конференции студентов, аспирантов и молодых ученых "Ломоносов - 2007"},
    language = {russian},
    publisher = {МГУ М.: МГУ},
    title = {Система управления компьютером на основе распознавания жестов лазерной указки},
    year = {2007}
}

Bibtex:

@inproceedings{jakubenko07trehmernaja,
    author = {Якубенко, А. and Конушин, А.},
    booktitle = {Материалы XIV Международной конференции студентов, аспирантов и молодых ученых "Ломоносов - 2007"},
    language = {russian},
    publisher = {МГУ М.: МГУ},
    title = {Трехмерная реконструкция городских зданий по изображениям},
    year = {2007}
}

Bibtex:

@inproceedings{grishin07bystryj,
    author = {Гришин, СВ and Ватолин, ДС},
    booktitle = {Материалы десятого научно-практи ческого семинара "Новые информационные технологии в автоматизиро ванных системах"},
    language = {russian},
    pages = {31--37},
    publisher = {Москва},
    title = {Быстрый алгоритм построения промежуточных кадров для многоракурсных стерео дисплеев},
    year = {2007}
}

Bibtex:

@inproceedings{grishin07metod,
    author = {Гришин, СВ and Ватолин, ДС},
    booktitle = {Труды 17-й Международной конференции по компьютерной графике и зрению ГрафиКон'2007},
    language = {russian},
    pages = {253--256},
    publisher = {Москва},
    series = {GraphiCon},
    title = {Метод повышения эффективности временной сегментации видео на сценах с медленным движением},
    year = {2007}
}

Bibtex:

@inproceedings{strelnikov07novyj,
    author = {Стрельников, Константин and Ватолин, Дмитрий and Солдатов, Станислав},
    booktitle = {Материалы десятого научно-практи ческого семинара "Новые информационные технологии в автоматизиро ванных системах"},
    language = {russian},
    pages = {3--10},
    publisher = {Москва},
    title = {Новый метод подавления эффекта дрожания кадра в видео},
    year = {2007}
}

Bibtex:

@inproceedings{senyukova07reconstruction,
    author = {Senyukova, O. and Yakubenko, A. and Degtiareva, A.},
    booktitle = {Proceedings of the 17th International Conference on Computer Graphics and Vision GraphiCon'2007},
    language = {english},
    pages = {219--222},
    publisher = {MAKS Press Moscow, Russia},
    series = {GraphiCon},
    title = {Reconstruction of generalized cylinders from two images},
    year = {2007}
}

Bibtex:

@inproceedings{senjukova07trehmernaja,
    author = {Сенюкова, О. and Дегтярева, А.},
    booktitle = {Материалы XIV Международной конференции студентов, аспирантов и молодых ученых "Ломоносов - 2007"},
    language = {russian},
    pages = {74},
    publisher = {МГУ М.: МГУ},
    title = {Трехмерная реконструкция объектов класса тел вращения по наброскам пользователя},
    year = {2007}
}

Bibtex:

@article{konouchine06extending,
    author = {Konouchine, A. S. and Gaganov, V. A. and Vezhnevets, V. P.},
    journal = {Sibirskii zhurnal vychislitel'noi matematiki},
    language = {english},
    number = {3},
    pages = {263--277},
    title = {Extending RANSAC-based estimators to handle unknown and varying noise level},
    volume = {9},
    year = {2006}
}

Bibtex:

@inproceedings{khropov06guided,
    author = {Khropov, A. and Konushin, A.},
    booktitle = {Proceedings of the 16th International Conference on Computer Graphics and Vision GraphiCon'2006},
    language = {english},
    pages = {47--53},
    publisher = {Novosibirsk Akademgorodok},
    series = {GraphiCon},
    title = {Guided Quasi-Dense Tracking for 3D Reconstruction},
    year = {2006}
}

Bibtex:

@inproceedings{yakubenko06imagebased,
    author = {Yakubenko, А. and Konouchine, А. and Vezhnevets, V.},
    booktitle = {Proceedings of the 16th International Conference on Computer Graphics and Vision GraphiCon'2006},
    language = {english},
    pages = {247--250},
    publisher = {Novosibirsk Akademgorodok},
    series = {GraphiCon},
    title = {Image-based 3D Reconstruction of Generalized Box with User Sketches},
    year = {2006}
}

Bibtex:

@inproceedings{lisitsin06podgonka,
    author = {Лисицин, Е. and Конушин, А.},
    booktitle = {Труды 16-й Международной конференции по компьютерной графике и зрению ГрафиКон'2006},
    language = {russian},
    pages = {371--374},
    publisher = {Россия, Новосибирск, Академгородок},
    series = {GraphiCon},
    title = {Подгонка параметрических моделей к облакам трехмерных точек},
    year = {2006}
}

Bibtex:

@inproceedings{vezhnevets06sovremennye,
    author = {Вежневец, В. and Конушин, А.},
    booktitle = {Тезисы докладов международной конференции "Тихонов и современная математика"},
    language = {russian},
    publisher = {МАКС Пресс Москва},
    title = {Современные методы прикладной математики в задачах обработки и распознавания изображений},
    year = {2006}
}

Bibtex:

@article{vatolin06cravnenija,
    author = {Ватолин, Дмитрий and Паршин, Александр},
    journal = {Matrix},
    language = {russian},
    number = {25},
    pages = {720--416},
    title = {Cравнения кодеков стандарта MPEG-4 AVC/H. 264 с использованием объективных метрик},
    volume = {239},
    year = {2006}
}

Abstract:

For practical applications there is a requirement of the global motion estimation that could use existing block matching motion data. In this paper authors propose two global motion estimation methods that are based on block matching motion estimation. These methods estimate different global motion models. One of the proposed method advantages consists in motion field filtration that allows coping with low reliability of motion vectors. In particular method of motion vectors reliability estimation is proposed. Model parameters are estimated sequentially, allowing to perform additional filtration to increase robustness. Proposed algorithms show good results in terms of robustness and precision while maintaining high processing speed and reasonable memory requirements.

Bibtex:

@inproceedings{soldatov06complexity,
    author = {Soldatov, Stanislav and Strelnikov, Konstantin and Vatolin, Dmitriy},
    booktitle = {Spring Conf. Computer Graphics},
    doi = {10.1.1.63.5502},
    language = {english},
    title = {Low complexity global motion estimation from block motion vectors},
    year = {2006}
}

Bibtex:

@inproceedings{vatolin06ntimes,
    author = {Vatolin, D. and Grishin, S.},
    booktitle = {Proceedings of the 16th International Conference on Computer Graphics and Vision GraphiCon'2006},
    language = {english},
    pages = {112--119},
    publisher = {Novosibirsk Akademgorodok},
    series = {GraphiCon},
    title = {N-times video frame-rate up-conversion algorithm based on pixel motion compensation with occlusions processing},
    year = {2006}
}

Bibtex:

@inproceedings{vatolin06objective,
    author = {Vatolin, Dmitriy and Grishin, Sergey and Moskvin, Alexey and Parshin, Alexander},
    booktitle = {SCCG},
    language = {english},
    title = {Objective Measurements of Artifacts, Produced by Modern Video Coding Standards},
    year = {2006}
}

Bibtex:

@inproceedings{grishin06algoritm,
    author = {Гришин, СВ and Ватолин, ДС},
    booktitle = {Труды 16-й Международной конференции по компьютерной графике и зрению ГрафиКон'2006},
    language = {russian},
    pages = {112--119},
    publisher = {Россия, Новосибирск, Академгородок},
    series = {GraphiCon},
    title = {Алгоритм N-кратного увеличения часто ты кадров видео на основе пиксельной компенсации движения с обработкой наложений},
    year = {2006}
}

Bibtex:

@inproceedings{vatolin06algoritm,
    author = {Ватолин, Дмитрий and Гришин, Сергей},
    booktitle = {Девятый научно-практический семинар ’Новые информационные технологии в автоматизированных системах’},
    language = {russian},
    pages = {32--46},
    title = {Алгоритм преобразования частоты кадров на основе интерполяции скомпенсированных кадров},
    year = {2006}
}

Bibtex:

@inproceedings{soldatov06bystroe,
    author = {Солдатов, СА and Стрельников, КН and Ватолин, ДС},
    booktitle = {Труды 16-й Международной конференции по компьютерной графике и зрению ГрафиКон'2006},
    language = {russian},
    pages = {430--437},
    publisher = {Россия, Новосибирск, Академгородок},
    series = {GraphiCon},
    title = {Быстрое и надежное определение глобального движения в видеопоследовательностях},
    year = {2006}
}

Bibtex:

@inproceedings{strelnikov06kachestvennoe,
    author = {Стрельников, КН and Солдатов, СА and Ватолин, ДС},
    booktitle = {материалы девятого научно-практического семинара "Новые информационные технологии в автоматизированных системах"},
    language = {russian},
    pages = {47--55},
    publisher = {Москва},
    title = {Качественное определение глобального движения кадра с использованием векторов движения},
    year = {2006}
}

Bibtex:

@inproceedings{vatolin06metody,
    author = {Ватолин, Дмитрий and Паршин, Александр},
    booktitle = {материалы девятого научно-практического семинара "Новые информационные технологии в автоматизированных системах"},
    language = {russian},
    pages = {4--12},
    publisher = {Москва},
    title = {Методы для объективной оценки качества видеокодеков по сжатым ими видеопоследовательностям},
    year = {2006}
}

Bibtex:

@inproceedings{mahinja06metody,
    author = {Махиня, М. Е. and Ватолин, Д. С.},
    booktitle = {Девятый научно-практический семинар ’Новые информационные технологии в автоматизированных системах’},
    language = {russian},
    pages = {56--58},
    title = {Методы пространственно-временного подавления шума в видео с использованием компенсации движения},
    volume = {9},
    year = {2006}
}

Bibtex:

@inproceedings{kulikov06obnaruzhenie,
    author = {Куликов, Дмитрий and Ватолин, Дмитрий},
    booktitle = {материалы девятого научно-практического семинара "Новые информационные технологии в автоматизированных системах"},
    language = {russian},
    pages = {13--23},
    publisher = {Москва},
    title = {Обнаружение и заполнение статических инородных областей в видео на примере удаления логотипов и сбоев при ошибках передачи},
    year = {2006}
}

Bibtex:

@inproceedings{kulikov06otsenka,
    author = {Куликов, ДЛ and Ватолин, ДС},
    booktitle = {Труды 16-й Международной конференции по компьютерной графике и зрению ГрафиКон'2006},
    language = {russian},
    pages = {1--4},
    publisher = {Россия, Новосибирск, Академгородок},
    series = {GraphiCon},
    title = {Оценка качества работы видео декодеров стандарта MPEG-2 при работе в ненадежной среде передачи данных},
    year = {2006}
}

Bibtex:

@inproceedings{degtiareva06reconstrustion,
    author = {Degtiareva, A. and Senyukova, O.},
    booktitle = {Proceedings of the 16th International Conference on Computer Graphics and Vision GraphiCon'2006},
    language = {english},
    pages = {308--311},
    publisher = {Novosibirsk Akademgorodok},
    series = {GraphiCon},
    title = {Reconstrustion of faces of revolution using two images},
    year = {2006}
}

Bibtex:

@article{senjukova06sistemy,
    author = {Сенюкова, О. and Дегтярева, А.},
    journal = {Компьютерная Графика и Мультимедиа - Сетевой журнал},
    language = {russian},
    title = {Системы моделирования трехмерных объектов},
    year = {2006}
}

Bibtex:

@inproceedings{konouchine05amlesac,
    author = {Konouchine, A. and Gaganov, V. and Veznevets, V.},
    booktitle = {Proceedings of the 15th International Conference on Computer Graphics and Vision GraphiCon'2005},
    language = {english},
    pages = {93--100},
    publisher = {Novosibirsk, Akademgorodok, Russia},
    series = {GraphiCon},
    title = {AMLESAC: A New Maximum Likelihood Robust Estimator},
    year = {2005}
}

Bibtex:

@inproceedings{konouchine05combined,
    author = {Konouchine, A. and Gaganov, V. and Veznevets, V.},
    booktitle = {Proceedings of the 15th International Conference on Computer Graphics and Vision GraphiCon'2005},
    language = {english},
    pages = {301--304},
    publisher = {Novosibirsk, Akademgorodok, Russia},
    series = {GraphiCon},
    title = {Combined Guided Tracking and Matching with Adaptive Track Initialization},
    year = {2005}
}

Bibtex:

@inproceedings{konushin05sistema,
    author = {Конушин, А. С.},
    booktitle = {Материалы Международной конференции студентов и аспирантов по фундаментальным наукам "Ломоносов 2005", секция "Вычислительная математика и кибернетика"},
    language = {russian},
    publisher = {издательство ВМК МГУ Москва},
    title = {Система построения трехмерных моделей реальных объектов по последовательности изображений},
    year = {2005}
}

Bibtex:

@article{kalinkina05problema,
    author = {Калинкина, Д. and Ватолин, Д.},
    journal = {Компьютерная Графика и Мультимедиа - Сетевой журнал},
    language = {russian},
    number = {2},
    title = {Проблема подавления шума на изображениях и видео и различные подходы к ее решению},
    volume = {9},
    year = {2005}
}

Bibtex:

@inproceedings{vatolin05evoljutsija,
    author = {Ватолин, Дмитрий},
    booktitle = {Proceedings of the 15th International Conference on Computer Graphics and Vision GraphiCon'2005},
    language = {russian},
    publisher = {Novosibirsk, Akademgorodok, Russia},
    series = {GraphiCon},
    title = {Эволюция стандартов сжатия видео},
    year = {2005}
}

Abstract:

We describe a novel depth image-based representation (DIBR) that has been adopted into MPEG-4 Animation Framework eXtension (AFX). The idea of this approach is to build a compact representation of a 3D object or scene without storing the geometry information in traditional polygonal form. This is done by using images accompanied by depth values (i.e, distance from image to the object) for each pixel. This type of representation allows to interactively build new views of the object (scene) according to the information contained in a few reference views. There are many different methods for the image-based rendering with depths, and DIBR format is designed to efficiently represent the information necessary for such methods. The main formats of the DIBR family SimpleTexture (an image together with depth array), PointTexture (a view of a scene from a single input camera but with multiple pixels along each line of sight), and OctreeImage (octree data structure together with a set of images containing viewport parameters). Detailed description of the OctreeImage compression methods used to compactly store and transmit the DIBR-objects is given. Rendering algorithms are outlined, and experimental results presented.

Bibtex:

@article{levkovichmaslyuk04depth,
    address = {Piscataway, NJ, United States},
    author = {Levkovich-Maslyuk, L. and Ignatenko, A. and Zhirkov, A. and Konushin, A. and Park, I. K. and Han, M. J. and Bayakovski, Y.},
    doi = {10.1109/TCSVT.2004.830676},
    issn = {1051-8215},
    journal = {IEEE Transactions on Circuits and Systems for Video Technology},
    language = {english},
    number = {7},
    pages = {1032--1045},
    publisher = {Piscataway, NJ, United States},
    title = {Depth image-based representation and compression for static and animated 3-D objects},
    volume = {14},
    year = {2004}
}

Bibtex:

@inproceedings{konushin04obzor,
    author = {Конушин, А. and Мариничев, К. and Вежневец, В.},
    booktitle = {Труды 14-й Международной конференции по компьютерной графике и зрению ГрафиКон'2004},
    language = {russian},
    pages = {275--278},
    publisher = {МГУ Москва},
    series = {GraphiCon},
    title = {Обзор робастных схем оценки параметров моделей на основе случайных выборок},
    year = {2004}
}

Bibtex:

@inproceedings{lisitsin04otslezhivanie,
    author = {Лисицин, E. and Конушин, А. and Вежневец, В.},
    booktitle = {Труды 14-й Международной конференции по компьютерной графике и зрению ГрафиКон'2004},
    language = {russian},
    pages = {233--236},
    publisher = {МГУ Москва},
    series = {GraphiCon},
    title = {Отслеживание точечных особенностей в видеопоследовательностях с изменениями резкости},
    year = {2004}
}

Bibtex:

@inproceedings{vatolin04kachestvennyj,
    author = {Ватолин, Дмитрий and Гришин, Сергей},
    booktitle = {Труды 14-й Международной конференции по компьютерной графике и зрению ГрафиКон'2004},
    language = {russian},
    pages = {257--260},
    publisher = {МГУ Москва},
    series = {GraphiCon},
    title = {Качественный метод деблокинга видео без использования информации о квантовании},
    year = {2004}
}

Bibtex:

@inproceedings{vatolin04effektivnyj,
    author = {Ватолин, Д. С. and Попов, Дмитрий},
    booktitle = {Proceedings of the 14th International Conference on Computer Graphics and Vision GraphiCon'2004},
    language = {russian},
    publisher = {Moscow, Russia},
    series = {GraphiCon},
    title = {Эффективный метод сжатия видео без потерь},
    year = {2004}
}

Bibtex:

@inproceedings{ignatenko03framework,
    author = {Ignatenko, A. and Konushin, A.},
    booktitle = {Proceedings of the 13th International Conference on Computer Graphics and Vision GraphiCon'2003},
    language = {english},
    pages = {169--172},
    publisher = {Moscow},
    series = {GraphiCon},
    title = {A Framework for Depth Image-Based Modeling and Rendering},
    year = {2003}
}

Bibtex:

@inproceedings{vatolin03otsenka,
    author = {Ватолин, Дмитрий and Путилин, Сергей},
    booktitle = {Труды 13-й Международной конференции по компьютерной графике и зрению ГрафиКон'2003},
    language = {russian},
    pages = {202--207},
    publisher = {Москва},
    series = {GraphiCon},
    title = {Оценка качества методов масштабирования изображений и результаты сравнений разных методов},
    year = {2003}
}

Bibtex:

@inproceedings{bayakovski02depth,
    author = {Bayakovski, Y. and Levkovich-Maslyuk, L. and Ignatenko, A. and Konushin, A. and Timasov, D. and Zhirkov, A. and Han, Mahnjin and Kyu, Park In},
    booktitle = {Proceedings of International Conference on Image Processing ICIP'02},
    doi = {10.1109/ICIP.2002.1038894},
    language = {english},
    pages = {25--28},
    title = {Depth image-based representations for static and animated 3D objects},
    volume = {3},
    year = {2002}
}

Abstract:

В книге описаны основные классические и современные методы сжатия: метод Хаффмана, арифметическое кодирование, LZ77, LZW, PPM, BWT, LPC и т. д. Разбираются алгоритмы, использующиеся в архиваторах Zip, HA, CabArc (*.cab-файлы), RAR, BZIP2, RK. Отдельный раздел посвящен алгоритмам сжатия изображений, использующимся в форматах PCX, TGA, GIF, TIFF, CCITT G-3, JPEG, JPEG2000. Рассмотрено фрактальное сжатие, вэйвлет-сжатие и др. Изложены принципы компрессии видеоданных, дан обзор стандартов MPEG, MPEG-2, MPEG-4, H.261 и H.263. Некоторые методы повышения сжатия на русском языке публикуются впервые. Книга содержит большое количество примеров и упражнений и ориентирована на студентов и преподавателей вузов. Материал книги позволяет самостоятельно несколькими способами написать архиватор с характеристиками, превосходящими программы типа pkzip и arj. Ответы на вопросы для самоконтроля и исходные тексты программ можно найти на сайте http://compression.graphicon.ru/.

Bibtex:

@book{vatolin02metody,
    author = {Ватолин, Д. and Ратушняк, А. and Смирнов, М. and Юкин, В.},
    isbn = {5-86404-170-X},
    language = {russian},
    numpages = {384},
    publisher = {Диалог-МИФИ Москва},
    title = {Методы сжатия данных. Устройство архиваторов, сжатие изображений и видео},
    year = {2002}
}

Bibtex:

@inproceedings{vatolin02primenenie,
    author = {Ватолин, Дмитрий},
    booktitle = {Труды 12-й Международной конференции по компьютерной графике и зрению ГрафиКон'2002},
    language = {russian},
    publisher = {Нижний Новгород},
    series = {GraphiCon},
    title = {Применение методов префильтрации видео для увеличения степени сжатия на потоках 600-1000 Кбит/с},
    year = {2002}
}

Bibtex:

@book{vatolin99algoritmy,
    author = {Ватолин, Д. С.},
    isbn = {5-89407-041-4},
    language = {russian},
    numpages = {76},
    publisher = {Издательский отдел факультета Вычислительной Математике и Кибернетики МГУ им. М.В. Ломоносова Москва},
    title = {Алгоритмы сжатия изображений. Москва},
    year = {1999}
}

Bibtex:

@article{vatolin99using,
    address = {Road Town, United Kingdom},
    author = {Vatolin, D. S.},
    issn = {1608-3261; 0361-7688},
    journal = {Programming and Computer Software},
    language = {english},
    number = {3},
    pages = {158--164},
    publisher = {Road Town, United Kingdom},
    title = {Using DCT for a fractal image compression optimization},
    volume = {25},
    year = {1999}
}

Bibtex:

@inproceedings{vatolin99visual,
    author = {Vatolin, Dmitriy},
    booktitle = {Proceedings of the 9th International Conference on Computer Graphics and Vision GraphiCon'1999},
    language = {english},
    series = {GraphiCon},
    title = {Visual Quality Increasing using Image Parts Improvement for Fractal Image Compression},
    year = {1999}
}

Bibtex:

@article{vatolin99ispolzovanie,
    author = {Ватолин, ДС},
    issn = {0132-3474},
    journal = {Программирование},
    language = {russian},
    number = {3},
    pages = {51--57},
    title = {Использование ДКП для ускорения фрактального сжатия изображений},
    year = {1999}
}

Bibtex:

@inproceedings{vatolin98hybrid,
    author = {Vatolin, Dmitriy},
    booktitle = {Proceedings of the 8th International Conference on Computer Graphics and Vision GraphiCon'1998},
    language = {english},
    series = {GraphiCon},
    title = {The hybrid method of a fractal image compression and vector quantization},
    year = {1998}
}

Bibtex:

@inproceedings{vatolin98gibridnaja,
    author = {Ватолин, Д. С.},
    booktitle = {Proceedings of the 8th International Conference on Computer Graphics and Vision GraphiCon'1998},
    language = {russian},
    series = {GraphiCon},
    title = {Гибридная схема фрактальной компрессии и квантования векторов для малых блоков},
    year = {1998}
}

Bibtex:

@article{vatolin96fraktalnoe,
    author = {Ватолин, ДС},
    journal = {ComputerWorld-Россия},
    language = {russian},
    number = {6},
    pages = {23},
    title = {Фрактальное сжатие изображений},
    year = {1996}
}

Bibtex:

@article{vatolin95mpegstandart,
    author = {Ватолин, Д. С.},
    journal = {Открытые системы},
    language = {russian},
    title = {MPEG-стандарт ISO на видео в системах мультимедиа},
    volume = {3},
    year = {1995}
}

Bibtex:

@article{vatolin95szhatie,
    author = {Ватолин, ДС},
    journal = {Открытые системы},
    language = {russian},
    number = {8},
    pages = {29},
    title = {Сжатие статических изображений},
    year = {1995}
}