{"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/multi-hierarchical-independent-correlation","title":"Multi-hierarchical Independent Correlation Filters for Visual Tracking","arxiv_id":"1811.10302","date":"2018-11-26","proceeding":null,"authors":["Shuai Bai","Zhiqun He","Ting-Bing Xu","Zheng Zhu","Yuan Dong","Hongliang Bai"],"abstract":"For visual tracking, most of the traditional correlation filters (CF) based\nmethods suffer from the bottleneck of feature redundancy and lack of motion\ninformation. In this paper, we design a novel tracking framework, called\nmulti-hierarchical independent correlation filters (MHIT). The framework\nconsists of motion estimation module, hierarchical features selection,\nindependent CF online learning, and adaptive multi-branch CF fusion.\nSpecifically, the motion estimation module is introduced to capture motion\ninformation, which effectively alleviates the object partial occlusion in the\ntemporal video. The multi-hierarchical deep features of CNN representing\ndifferent semantic information can be fully excavated to track multi-scale\nobjects. To better overcome the deep feature redundancy, each hierarchical\nfeatures are independently fed into a single branch to implement the online\nlearning of parameters. Finally, an adaptive weight scheme is integrated into\nthe framework to fuse these independent multi-branch CFs for the better and\nmore robust visual object tracking. Extensive experiments on OTB and VOT\ndatasets show that the proposed MHIT tracker can significantly improve the\ntracking performance. Especially, it obtains a 20.1% relative performance gain\ncompared to the top trackers on the VOT2017 challenge, and also achieves new\nstate-of-the-art performance on the VOT2018 challenge.","url_abs":"http://arxiv.org/abs/1811.10302v2","url_pdf":"http://arxiv.org/pdf/1811.10302v2.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":"multi-hierarchical-independent-correlation","repo_url":"https://github.com/ShuaiBai623/MFT","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"motion-estimation","task_name":"Motion Estimation"},{"task_slug":"object-tracking","task_name":"Object Tracking"},{"task_slug":"visual-object-tracking","task_name":"Visual Object Tracking"},{"task_slug":"visual-tracking","task_name":"Visual Tracking"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1811.10302","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}