{"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/generative-partition-networks-for-multi","title":"Generative Partition Networks for Multi-Person Pose Estimation","arxiv_id":"1705.07422","date":"2017-05-21","proceeding":null,"authors":["Xuecheng Nie","Jiashi Feng","Junliang Xing","Shuicheng Yan"],"abstract":"This paper proposes a new Generative Partition Network (GPN) to address the\nchallenging multi-person pose estimation problem. Different from existing\nmodels that are either completely top-down or bottom-up, the proposed GPN\nintroduces a novel strategy--it generates partitions for multiple persons from\ntheir global joint candidates and infers instance-specific joint configurations\nsimultaneously. The GPN is favorably featured by low complexity and high\naccuracy of joint detection and re-organization. In particular, GPN designs a\ngenerative model that performs one feed-forward pass to efficiently generate\nrobust person detections with joint partitions, relying on dense regressions\nfrom global joint candidates in an embedding space parameterized by centroids\nof persons. In addition, GPN formulates the inference procedure for joint\nconfigurations of human poses as a graph partition problem, and conducts local\noptimization for each person detection with reliable global affinity cues,\nleading to complexity reduction and performance improvement. GPN is implemented\nwith the Hourglass architecture as the backbone network to simultaneously learn\njoint detector and dense regressor. Extensive experiments on benchmarks MPII\nHuman Pose Multi-Person, extended PASCAL-Person-Part, and WAF, show the\nefficiency of GPN with new state-of-the-art performance.","url_abs":"http://arxiv.org/abs/1705.07422v2","url_pdf":"http://arxiv.org/pdf/1705.07422v2.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":"generative-partition-networks-for-multi","repo_url":"https://github.com/NieXC/pytorch-ppn","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"human-detection","task_name":"Human Detection"},{"task_slug":"keypoint-detection","task_name":"Keypoint Detection"},{"task_slug":"multi-person-pose-estimation","task_name":"Multi-Person Pose Estimation"},{"task_slug":"pose-estimation","task_name":"Pose Estimation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/keypoint-detection-on-mpii-multi-person","task":"Keypoint Detection","dataset":"MPII Multi-Person","model":"Generative Partition Networks","rank_in_archive_order":2,"of":9,"metrics":{"mAP@0.5":"80.4%"},"uses_additional_data":false},{"leaderboard":"/sota/multi-person-pose-estimation-on-mpii-multi","task":"Multi-Person Pose Estimation","dataset":"MPII Multi-Person","model":"Generative Partition Networks","rank_in_archive_order":2,"of":9,"metrics":{"AP":"80.4%"},"uses_additional_data":false},{"leaderboard":"/sota/multi-person-pose-estimation-on-waf","task":"Multi-Person Pose Estimation","dataset":"WAF","model":"Generative Partition Networks","rank_in_archive_order":3,"of":3,"metrics":{"AP":"84.8"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1705.07422","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}