{"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/repetition-estimation","title":"Repetition Estimation","arxiv_id":"1806.06984","date":"2018-06-18","proceeding":null,"authors":["Tom F. H. Runia","Cees G. M. Snoek","Arnold W. M. Smeulders"],"abstract":"Visual repetition is ubiquitous in our world. It appears in human activity\n(sports, cooking), animal behavior (a bee's waggle dance), natural phenomena\n(leaves in the wind) and in urban environments (flashing lights). Estimating\nvisual repetition from realistic video is challenging as periodic motion is\nrarely perfectly static and stationary. To better deal with realistic video, we\nelevate the static and stationary assumptions often made by existing work. Our\nspatiotemporal filtering approach, established on the theory of periodic\nmotion, effectively handles a wide variety of appearances and requires no\nlearning. Starting from motion in 3D we derive three periodic motion types by\ndecomposition of the motion field into its fundamental components. In addition,\nthree temporal motion continuities emerge from the field's temporal dynamics.\nFor the 2D perception of 3D motion we consider the viewpoint relative to the\nmotion; what follows are 18 cases of recurrent motion perception. To estimate\nrepetition under all circumstances, our theory implies constructing a mixture\nof differential motion maps: gradient, divergence and curl. We temporally\nconvolve the motion maps with wavelet filters to estimate repetitive dynamics.\nOur method is able to spatially segment repetitive motion directly from the\ntemporal filter responses densely computed over the motion maps. For\nexperimental verification of our claims, we use our novel dataset for\nrepetition estimation, better-reflecting reality with non-static and\nnon-stationary repetitive motion. On the task of repetition counting, we obtain\nfavorable results compared to a deep learning alternative.","url_abs":"http://arxiv.org/abs/1806.06984v1","url_pdf":"http://arxiv.org/pdf/1806.06984v1.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":"repetition-estimation","repo_url":"https://github.com/tomrunia/PyTorchWavelets","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}