{"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/switchable-temporal-propagation-network","title":"Switchable Temporal Propagation Network","arxiv_id":"1804.08758","date":"2018-04-23","proceeding":"ECCV 2018 9","authors":["Sifei Liu","Guangyu Zhong","Shalini De Mello","Jinwei Gu","Varun Jampani","Ming-Hsuan Yang","Jan Kautz"],"abstract":"Videos contain highly redundant information between frames. Such redundancy\nhas been extensively studied in video compression and encoding, but is less\nexplored for more advanced video processing. In this paper, we propose a\nlearnable unified framework for propagating a variety of visual properties of\nvideo images, including but not limited to color, high dynamic range (HDR), and\nsegmentation information, where the properties are available for only a few\nkey-frames. Our approach is based on a temporal propagation network (TPN),\nwhich models the transition-related affinity between a pair of frames in a\npurely data-driven manner. We theoretically prove two essential factors for\nTPN: (a) by regularizing the global transformation matrix as orthogonal, the\n\"style energy\" of the property can be well preserved during propagation; (b)\nsuch regularization can be achieved by the proposed switchable TPN with\nbi-directional training on pairs of frames. We apply the switchable TPN to\nthree tasks: colorizing a gray-scale video based on a few color key-frames,\ngenerating an HDR video from a low dynamic range (LDR) video and a few HDR\nframes, and propagating a segmentation mask from the first frame in videos.\nExperimental results show that our approach is significantly more accurate and\nefficient than the state-of-the-art methods.","url_abs":"http://arxiv.org/abs/1804.08758v2","url_pdf":"http://arxiv.org/pdf/1804.08758v2.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":"switchable-temporal-propagation-network","repo_url":"https://github.com/Liusifei/UVC","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"video-compression","task_name":"Video Compression"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}