{"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/dynamic-label-graph-matching-for-unsupervised","title":"Dynamic Label Graph Matching for Unsupervised Video Re-Identification","arxiv_id":"1709.09297","date":"2017-09-27","proceeding":"ICCV 2017 10","authors":["Mang Ye","Andy J. Ma","Liang Zheng","Jiawei Li","P C Yuen"],"abstract":"Label estimation is an important component in an unsupervised person\nre-identification (re-ID) system. This paper focuses on cross-camera label\nestimation, which can be subsequently used in feature learning to learn robust\nre-ID models. Specifically, we propose to construct a graph for samples in each\ncamera, and then graph matching scheme is introduced for cross-camera labeling\nassociation. While labels directly output from existing graph matching methods\nmay be noisy and inaccurate due to significant cross-camera variations, this\npaper proposes a dynamic graph matching (DGM) method. DGM iteratively updates\nthe image graph and the label estimation process by learning a better feature\nspace with intermediate estimated labels. DGM is advantageous in two aspects:\n1) the accuracy of estimated labels is improved significantly with the\niterations; 2) DGM is robust to noisy initial training data. Extensive\nexperiments conducted on three benchmarks including the large-scale MARS\ndataset show that DGM yields competitive performance to fully supervised\nbaselines, and outperforms competing unsupervised learning methods.","url_abs":"http://arxiv.org/abs/1709.09297v1","url_pdf":"http://arxiv.org/pdf/1709.09297v1.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":"graph-matching","task_name":"Graph Matching"},{"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":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/person-re-identification-on-prid2011","task":"Person Re-Identification","dataset":"PRID2011","model":"DGM+MLAPG+","rank_in_archive_order":8,"of":13,"metrics":{"Rank-1":"73.1","Rank-20":"99.0","Rank-5":"92.5"},"uses_additional_data":false},{"leaderboard":"/sota/person-re-identification-on-prid2011","task":"Person Re-Identification","dataset":"PRID2011","model":"DGM+IDE+","rank_in_archive_order":9,"of":13,"metrics":{"Rank-1":"56.4","Rank-20":"96.4","Rank-5":"81.3"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}