{"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/near-online-multi-target-tracking-with","title":"Near-Online Multi-target Tracking with Aggregated Local Flow Descriptor","arxiv_id":"1504.02340","date":"2015-04-09","proceeding":"ICCV 2015 12","authors":["Wongun Choi"],"abstract":"In this paper, we focus on the two key aspects of multiple target tracking\nproblem: 1) designing an accurate affinity measure to associate detections and\n2) implementing an efficient and accurate (near) online multiple target\ntracking algorithm. As the first contribution, we introduce a novel Aggregated\nLocal Flow Descriptor (ALFD) that encodes the relative motion pattern between a\npair of temporally distant detections using long term interest point\ntrajectories (IPTs). Leveraging on the IPTs, the ALFD provides a robust\naffinity measure for estimating the likelihood of matching detections\nregardless of the application scenarios. As another contribution, we present a\nNear-Online Multi-target Tracking (NOMT) algorithm. The tracking problem is\nformulated as a data-association between targets and detections in a temporal\nwindow, that is performed repeatedly at every frame. While being efficient,\nNOMT achieves robustness via integrating multiple cues including ALFD metric,\ntarget dynamics, appearance similarity, and long term trajectory regularization\ninto the model. Our ablative analysis verifies the superiority of the ALFD\nmetric over the other conventional affinity metrics. We run a comprehensive\nexperimental evaluation on two challenging tracking datasets, KITTI and MOT\ndatasets. The NOMT method combined with ALFD metric achieves the best accuracy\nin both datasets with significant margins (about 10% higher MOTA) over the\nstate-of-the-arts.","url_abs":"http://arxiv.org/abs/1504.02340v1","url_pdf":"http://arxiv.org/pdf/1504.02340v1.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":[],"tasks":[{"task_slug":"multiple-object-tracking","task_name":"Multiple Object Tracking"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/multi-object-tracking-on-mot16","task":"Multi-Object Tracking","dataset":"MOT16","model":"NOMT","rank_in_archive_order":23,"of":24,"metrics":{"MOTA":"46.4"},"uses_additional_data":false},{"leaderboard":"/sota/multiple-object-tracking-on-kitti-test-online","task":"Multiple Object Tracking","dataset":"KITTI Test (Online Methods)","model":"NOMT","rank_in_archive_order":28,"of":34,"metrics":{"MOTA":"78.15"},"uses_additional_data":false},{"leaderboard":"/sota/multiple-object-tracking-on-kitti-test-online","task":"Multiple Object Tracking","dataset":"KITTI Test (Online Methods)","model":"NOMT-HM","rank_in_archive_order":33,"of":34,"metrics":{"MOTA":"75.20"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1504.02340","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}