{"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/beyond-correlation-filters-learning","title":"Beyond Correlation Filters: Learning Continuous Convolution Operators for Visual Tracking","arxiv_id":"1608.03773","date":"2016-08-12","proceeding":null,"authors":["Martin Danelljan","Andreas Robinson","Fahad Shahbaz Khan","Michael Felsberg"],"abstract":"Discriminative Correlation Filters (DCF) have demonstrated excellent\nperformance for visual object tracking. The key to their success is the ability\nto efficiently exploit available negative data by including all shifted\nversions of a training sample. However, the underlying DCF formulation is\nrestricted to single-resolution feature maps, significantly limiting its\npotential. In this paper, we go beyond the conventional DCF framework and\nintroduce a novel formulation for training continuous convolution filters. We\nemploy an implicit interpolation model to pose the learning problem in the\ncontinuous spatial domain. Our proposed formulation enables efficient\nintegration of multi-resolution deep feature maps, leading to superior results\non three object tracking benchmarks: OTB-2015 (+5.1% in mean OP), Temple-Color\n(+4.6% in mean OP), and VOT2015 (20% relative reduction in failure rate).\nAdditionally, our approach is capable of sub-pixel localization, crucial for\nthe task of accurate feature point tracking. We also demonstrate the\neffectiveness of our learning formulation in extensive feature point tracking\nexperiments. Code and supplementary material are available at\nhttp://www.cvl.isy.liu.se/research/objrec/visualtracking/conttrack/index.html.","url_abs":"http://arxiv.org/abs/1608.03773v2","url_pdf":"http://arxiv.org/pdf/1608.03773v2.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":"beyond-correlation-filters-learning","repo_url":"https://github.com/martin-danelljan/Continuous-ConvOp","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"GPL-3.0"}}],"tasks":[{"task_slug":"object-tracking","task_name":"Object Tracking"},{"task_slug":"point-tracking","task_name":"Point Tracking"},{"task_slug":"visual-object-tracking","task_name":"Visual Object Tracking"},{"task_slug":"visual-tracking","task_name":"Visual Tracking"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1608.03773","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}