{"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/online-multi-object-tracking-with-dual","title":"Online Multi-Object Tracking with Dual Matching Attention Networks","arxiv_id":"1902.00749","date":"2019-02-02","proceeding":"ECCV 2018 9","authors":["Ji Zhu","Hua Yang","Nian Liu","Minyoung Kim","Wenjun Zhang","Ming-Hsuan Yang"],"abstract":"In this paper, we propose an online Multi-Object Tracking (MOT) approach\nwhich integrates the merits of single object tracking and data association\nmethods in a unified framework to handle noisy detections and frequent\ninteractions between targets. Specifically, for applying single object tracking\nin MOT, we introduce a cost-sensitive tracking loss based on the\nstate-of-the-art visual tracker, which encourages the model to focus on hard\nnegative distractors during online learning. For data association, we propose\nDual Matching Attention Networks (DMAN) with both spatial and temporal\nattention mechanisms. The spatial attention module generates dual attention\nmaps which enable the network to focus on the matching patterns of the input\nimage pair, while the temporal attention module adaptively allocates different\nlevels of attention to different samples in the tracklet to suppress noisy\nobservations. Experimental results on the MOT benchmark datasets show that the\nproposed algorithm performs favorably against both online and offline trackers\nin terms of identity-preserving metrics.","url_abs":"http://arxiv.org/abs/1902.00749v1","url_pdf":"http://arxiv.org/pdf/1902.00749v1.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":"online-multi-object-tracking-with-dual","repo_url":"https://github.com/jizhu1023/DMAN_MOT","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"multi-object-tracking","task_name":"Multi-Object Tracking"},{"task_slug":"object","task_name":"Object"},{"task_slug":"object-tracking","task_name":"Object Tracking"},{"task_slug":"online-multi-object-tracking","task_name":"Online Multi-Object Tracking"}],"methods":[{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/multi-object-tracking-on-mot16","task":"Multi-Object Tracking","dataset":"MOT16","model":"DMMOT","rank_in_archive_order":24,"of":24,"metrics":{"MOTA":"46.1"},"uses_additional_data":false},{"leaderboard":"/sota/online-multi-object-tracking-on-mot16","task":"Online Multi-Object Tracking","dataset":"MOT16","model":"DMAN","rank_in_archive_order":5,"of":5,"metrics":{"MOTA":"46.1"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1902.00749","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}