{"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/unsupervised-person-re-identification-by-deep-1","title":"Unsupervised Person Re-identification by Deep Learning Tracklet Association","arxiv_id":"1809.02874","date":"2018-09-08","proceeding":"ECCV 2018 9","authors":["Minxian Li","Xiatian Zhu","Shaogang Gong"],"abstract":"Mostexistingpersonre-identification(re-id)methods relyon supervised model\nlearning on per-camera-pair manually labelled pairwise training data. This\nleads to poor scalability in practical re-id deployment due to the lack of\nexhaustive identity labelling of image positive and negative pairs for every\ncamera pair. In this work, we address this problem by proposing an unsupervised\nre-id deep learning approach capable of incrementally discovering and\nexploiting the underlying re-id discriminative information from automatically\ngenerated person tracklet data from videos in an end-to-end model optimisation.\nWe formulate a Tracklet Association Unsupervised Deep Learning (TAUDL)\nframework characterised by jointly learning per-camera (within-camera) tracklet\nassociation (labelling) and cross-camera tracklet correlation by maximising the\ndiscovery of most likely tracklet relationships across camera views. Extensive\nexperiments demonstrate the superiority of the proposed TAUDL model over the\nstate-of-the-art unsupervised and domain adaptation re- id methods using six\nperson re-id benchmarking datasets.","url_abs":"http://arxiv.org/abs/1809.02874v1","url_pdf":"http://arxiv.org/pdf/1809.02874v1.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":"benchmarking","task_name":"Benchmarking"},{"task_slug":"deep-learning","task_name":"Deep Learning"},{"task_slug":"domain-adaptation","task_name":"Domain Adaptation"},{"task_slug":"person-re-identification","task_name":"Person Re-Identification"},{"task_slug":"unsupervised-person-re-identification","task_name":"Unsupervised Person Re-Identification"}],"methods":[],"datasets_introduced":[{"slug":"ilids-vid","name":"iLIDS-VID","full_name":"iLIDS-VID"}],"methods_introduced":[],"results":[{"leaderboard":"/sota/person-re-identification-on-duketracklet","task":"Person Re-Identification","dataset":"DukeTracklet","model":"TAUDL","rank_in_archive_order":2,"of":2,"metrics":{"Rank-1":"26.1","Rank-20":"57.2","Rank-5":"42.0","mAP":"20.8"},"uses_additional_data":false},{"leaderboard":"/sota/person-re-identification-on-msmt17","task":"Person Re-Identification","dataset":"MSMT17","model":"TAUDL","rank_in_archive_order":42,"of":43,"metrics":{"mAP":"12.5"},"uses_additional_data":false},{"leaderboard":"/sota/person-re-identification-on-prid2011","task":"Person Re-Identification","dataset":"PRID2011","model":"TAUDL","rank_in_archive_order":12,"of":13,"metrics":{"Rank-1":"49.4","Rank-20":"98.9","Rank-5":"78.7"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1809.02874","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}