{"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/using-phase-instead-of-optical-flow-for","title":"Using phase instead of optical flow for action recognition","arxiv_id":"1809.03258","date":"2018-09-10","proceeding":null,"authors":["Omar Hommos","Silvia L. Pintea","Pascal S. M. Mettes","Jan C. van Gemert"],"abstract":"Currently, the most common motion representation for action recognition is\noptical flow. Optical flow is based on particle tracking which adheres to a\nLagrangian perspective on dynamics. In contrast to the Lagrangian perspective,\nthe Eulerian model of dynamics does not track, but describes local changes. For\nvideo, an Eulerian phase-based motion representation, using complex steerable\nfilters, has been successfully employed recently for motion magnification and\nvideo frame interpolation. Inspired by these previous works, here, we proposes\nlearning Eulerian motion representations in a deep architecture for action\nrecognition. We learn filters in the complex domain in an end-to-end manner. We\ndesign these complex filters to resemble complex Gabor filters, typically\nemployed for phase-information extraction. We propose a phase-information\nextraction module, based on these complex filters, that can be used in any\nnetwork architecture for extracting Eulerian representations. We experimentally\nanalyze the added value of Eulerian motion representations, as extracted by our\nproposed phase extraction module, and compare with existing motion\nrepresentations based on optical flow, on the UCF101 dataset.","url_abs":"http://arxiv.org/abs/1809.03258v2","url_pdf":"http://arxiv.org/pdf/1809.03258v2.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":"using-phase-instead-of-optical-flow-for","repo_url":"https://github.com/11maxed11/phase-based-action-recognition","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"action-recognition-in-videos","task_name":"Action Recognition"},{"task_slug":"motion-magnification","task_name":"Motion Magnification"},{"task_slug":"optical-flow-estimation","task_name":"Optical Flow Estimation"},{"task_slug":"action-recognition","task_name":"Temporal Action Localization"},{"task_slug":"video-frame-interpolation","task_name":"Video Frame Interpolation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1809.03258","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}