%0 Conference Paper %1 6751550 %A Galasso, F. %A Nagaraja, N.S. %A Jimenez Cardenas, T. %A Brox, T. %A Schiele, B. %B Computer Vision (ICCV), 2013 IEEE International Conference on %D 2013 %K ComputerVision motion_analysis motion_segmentation segmentation %P 3527-3534 %R 10.1109/ICCV.2013.438 %T A Unified Video Segmentation Benchmark: Annotation, Metrics and Analysis %U http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=6751550&tag=1 %X Video segmentation research is currently limited by the lack of a benchmark dataset that covers the large variety of sub problems appearing in video segmentation and that is large enough to avoid over fitting. Consequently, there is little analysis of video segmentation which generalizes across subtasks, and it is not yet clear which and how video segmentation should leverage the information from the still-frames, as previously studied in image segmentation, alongside video specific information, such as temporal volume, motion and occlusion. In this work we provide such an analysis based on annotations of a large video dataset, where each video is manually segmented by multiple persons. Moreover, we introduce a new volume-based metric that includes the important aspect of temporal consistency, that can deal with segmentation hierarchies, and that reflects the tradeoff between over-segmentation and segmentation accuracy.

@inproceedings{6751550, abstract = {Video segmentation research is currently limited by the lack of a benchmark dataset that covers the large variety of sub problems appearing in video segmentation and that is large enough to avoid over fitting. Consequently, there is little analysis of video segmentation which generalizes across subtasks, and it is not yet clear which and how video segmentation should leverage the information from the still-frames, as previously studied in image segmentation, alongside video specific information, such as temporal volume, motion and occlusion. In this work we provide such an analysis based on annotations of a large video dataset, where each video is manually segmented by multiple persons. Moreover, we introduce a new volume-based metric that includes the important aspect of temporal consistency, that can deal with segmentation hierarchies, and that reflects the tradeoff between over-segmentation and segmentation accuracy.}, added-at = {2014-07-29T13:45:53.000+0200}, author = {Galasso, F. and Nagaraja, N.S. and Jimenez Cardenas, T. and Brox, T. and Schiele, B.}, biburl = {https://www.bibsonomy.org/bibtex/237ed30dfb13b3a175fbe2021d77e4cb0/alex_ruff}, booktitle = {Computer Vision (ICCV), 2013 IEEE International Conference on}, description = {IEEE Xplore Abstract - A Unified Video Segmentation Benchmark: Annotation, Metrics and Analysis}, doi = {10.1109/ICCV.2013.438}, interhash = {fda340e6c9bcdd1ec79709bc9866f1e7}, intrahash = {37ed30dfb13b3a175fbe2021d77e4cb0}, issn = {1550-5499}, keywords = {ComputerVision motion_analysis motion_segmentation segmentation}, month = dec, pages = {3527-3534}, timestamp = {2014-07-29T13:45:53.000+0200}, title = {A Unified Video Segmentation Benchmark: Annotation, Metrics and Analysis}, url = {http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=6751550&tag=1}, year = 2013 }