{"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/exciting-inhibition-network-for-person","title":"Exciting-Inhibition Network for Person Reidentification in Internet of Things","arxiv_id":null,"date":"2021-10-15","proceeding":"IEEE Internet of Things Journal 2021 10","authors":["Meixia Fu","Songlin Sun","Qilian Liang","Xiaoyun Tong","Qiang Liu"],"abstract":"Person reidentification (re-ID), which aims at rec\u0002ognizing the pedestrians captured by multiple nonoverlapping\r\ncameras, has attracted more interest due to its significant and\r\npotential application in the Internet of Things like intelligent\r\nvisual surveillance. However, person reID is still a challenging\r\nproblem in the situations of various pose, similar appearances,\r\npartial occlusion, etc. To handle these obstacles, in this article, we\r\ninvestigate an innovative exciting-inhibition network (EINet) that\r\nis a two-branch network composed of the exciting branch and\r\nthe inhibition branch. The channel-spatial attention block that\r\nrecalibrates the relationship between channels and highlights fea\u0002tures at different spatial positions is used in the exciting branch.\r\nA novel Soft Batch DropBlock that randomly selects a continu\u0002ous region of the intermediate feature maps at the same location\r\nis applied in the inhibition branch to inhibit the trivial by an\r\ninhibitive mask and reinforce learning the remaining regions.\r\nWe integrate the comprehensive features from both branches\r\nfor evaluation and show the performance of EINet intuitively\r\nusing the visualization method. Abundant experiments demon\u0002strate the state-of-the-art performance by comparing with the\r\nprevious methods on three popular person re-ID benchmarks. For\r\nexample, our method obtains 95.64% Rank-1 and 88.75% mean\r\naverage precision (mAP) on Market-1501, and 77.00% Rank-1\r\nand 74.51% mAP on CUHK03-Detect in the single query mode,\r\nrespectively.\r\nIndex Terms—Channel-spatial attention block (CSAB),\r\nexciting-inhibition network (EINet), Internet of Things (IoT),\r\nperson reidentification (re-ID), soft batch dropblock.","url_abs":"https://ieeexploreieee.53yu.com/abstract/document/9252840/","url_pdf":"https://ieeexploreieee.53yu.com/abstract/document/9252840","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":"person-re-identification","task_name":"Person Re-Identification"}],"methods":[{"method_slug":"dropblock","method_name":"DropBlock"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/person-re-identification-on-market-1501","task":"Person Re-Identification","dataset":"Market-1501","model":"EI-Net","rank_in_archive_order":130,"of":135,"metrics":{"mAP":"88.75"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}