{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/motion-segmentation-by-exploiting","title":"Motion Segmentation by Exploiting Complementary Geometric Models","arxiv_id":"1804.02142","date":"2018-04-06","proceeding":"CVPR 2018 6","authors":["Xun Xu","Loong-Fah Cheong","Zhuwen Li"],"abstract":"Many real-world sequences cannot be conveniently categorized as general or\ndegenerate; in such cases, imposing a false dichotomy in using the fundamental\nmatrix or homography model for motion segmentation would lead to difficulty.\nEven when we are confronted with a general scene-motion, the fundamental matrix\napproach as a model for motion segmentation still suffers from several defects,\nwhich we discuss in this paper. The full potential of the fundamental matrix\napproach could only be realized if we judiciously harness information from the\nsimpler homography model. From these considerations, we propose a multi-view\nspectral clustering framework that synergistically combines multiple models\ntogether. We show that the performance can be substantially improved in this\nway. We perform extensive testing on existing motion segmentation datasets,\nachieving state-of-the-art performance on all of them; we also put forth a more\nrealistic and challenging dataset adapted from the KITTI benchmark, containing\nreal-world effects such as strong perspectives and strong forward translations\nnot seen in the traditional datasets.","url_abs":"http://arxiv.org/abs/1804.02142v1","url_pdf":"http://arxiv.org/pdf/1804.02142v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"motion-segmentation-by-exploiting","repo_url":"https://github.com/alex-xun-xu/MultiViewMoSeg","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"clustering","task_name":"Clustering"},{"task_slug":"motion-segmentation","task_name":"Motion Segmentation"},{"task_slug":"segmentation","task_name":"Segmentation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/motion-segmentation-on-hopkins155","task":"Motion Segmentation","dataset":"Hopkins155","model":"MVC","rank_in_archive_order":1,"of":4,"metrics":{"Classification Error":"0.31"},"uses_additional_data":false},{"leaderboard":"/sota/motion-segmentation-on-kt3dmoseg","task":"Motion Segmentation","dataset":"KT3DMoSeg","model":"MultiViewClustering","rank_in_archive_order":1,"of":1,"metrics":{"Error":"7.92"},"uses_additional_data":false},{"leaderboard":"/sota/motion-segmentation-on-mtpv62","task":"Motion Segmentation","dataset":"MTPV62","model":"MVC","rank_in_archive_order":1,"of":1,"metrics":{"Classification Error":"0.65"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1804.02142","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}