{"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/improved-person-re-identification-based-on","title":"Improved Person Re-Identification Based on Saliency and Semantic Parsing with Deep Neural Network Models","arxiv_id":"1807.05618","date":"2018-07-15","proceeding":null,"authors":["Rodolfo Quispe","Helio Pedrini"],"abstract":"Given a video or an image of a person acquired from a camera, person\nre-identification is the process of retrieving all instances of the same person\nfrom videos or images taken from a different camera with non-overlapping view.\nThis task has applications in various fields, such as surveillance, forensics,\nrobotics, multimedia. In this paper, we present a novel framework, named\nSaliency-Semantic Parsing Re-Identification (SSP-ReID), for taking advantage of\nthe capabilities of both clues: saliency and semantic parsing maps, to guide a\nbackbone convolutional neural network (CNN) to learn complementary\nrepresentations that improves the results over the original backbones. The\ninsight of fusing multiple clues is based on specific scenarios in which one\nresponse is better than another, thus favoring the combination of them to\nincrease performance. Due to its definition, our framework can be easily\napplied to a wide variety of networks and, in contrast to other competitive\nmethods, our training process follows simple and standard protocols. We present\nextensive evaluation of our approach through five backbones and three\nbenchmarks. Experimental results demonstrate the effectiveness of our person\nre-identification framework. In addition, we combine our framework with\nre-ranking techniques to achieve state-of-the-art results on three benchmarks.","url_abs":"http://arxiv.org/abs/1807.05618v1","url_pdf":"http://arxiv.org/pdf/1807.05618v1.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":"improved-person-re-identification-based-on","repo_url":"https://github.com/RQuispeC/saliency-semantic-parsing-reid","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"person-re-identification","task_name":"Person Re-Identification"},{"task_slug":"re-ranking","task_name":"Re-Ranking"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/person-re-identification-on-dukemtmc-reid","task":"Person Re-Identification","dataset":"DukeMTMC-reID","model":"SSP-ReID (RR)","rank_in_archive_order":26,"of":94,"metrics":{"Rank-1":"86.4","mAP":"83.7"},"uses_additional_data":false},{"leaderboard":"/sota/person-re-identification-on-dukemtmc-reid","task":"Person Re-Identification","dataset":"DukeMTMC-reID","model":"SSP-ReID","rank_in_archive_order":66,"of":94,"metrics":{"Rank-1":"81.8","mAP":"68.6"},"uses_additional_data":false},{"leaderboard":"/sota/person-re-identification-on-market-1501","task":"Person Re-Identification","dataset":"Market-1501","model":"SSP-ReID (RR)","rank_in_archive_order":83,"of":135,"metrics":{"Rank-1":"93.7","mAP":"90.8"},"uses_additional_data":false},{"leaderboard":"/sota/person-re-identification-on-market-1501","task":"Person Re-Identification","dataset":"Market-1501","model":"SSP-ReID","rank_in_archive_order":88,"of":135,"metrics":{"Rank-1":"92.5","mAP":"80.1"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}