{"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/joint-unsupervised-learning-of-optical-flow","title":"Joint Unsupervised Learning of Optical Flow and Depth by Watching Stereo Videos","arxiv_id":"1810.03654","date":"2018-10-08","proceeding":null,"authors":["Yang Wang","Zhenheng Yang","Peng Wang","Yi Yang","Chenxu Luo","Wei Xu"],"abstract":"Learning depth and optical flow via deep neural networks by watching videos\nhas made significant progress recently. In this paper, we jointly solve the two\ntasks by exploiting the underlying geometric rules within stereo videos.\nSpecifically, given two consecutive stereo image pairs from a video, we first\nestimate depth, camera ego-motion and optical flow from three neural networks.\nThen the whole scene is decomposed into moving foreground and static background\nby compar- ing the estimated optical flow and rigid flow derived from the depth\nand ego-motion. We propose a novel consistency loss to let the optical flow\nlearn from the more accurate rigid flow in static regions. We also design a\nrigid alignment module which helps refine ego-motion estimation by using the\nestimated depth and optical flow. Experiments on the KITTI dataset show that\nour results significantly outperform other state-of- the-art algorithms. Source\ncodes can be found at https: //github.com/baidu-research/UnDepthflow","url_abs":"http://arxiv.org/abs/1810.03654v1","url_pdf":"http://arxiv.org/pdf/1810.03654v1.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":"joint-unsupervised-learning-of-optical-flow","repo_url":"https://github.com/baidu-research/UnDepthflow","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"motion-estimation","task_name":"Motion Estimation"},{"task_slug":"optical-flow-estimation","task_name":"Optical Flow Estimation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}