{"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/posetrack-a-benchmark-for-human-pose","title":"PoseTrack: A Benchmark for Human Pose Estimation and Tracking","arxiv_id":"1710.10000","date":"2017-10-27","proceeding":"CVPR 2018 6","authors":["Mykhaylo Andriluka","Umar Iqbal","Eldar Insafutdinov","Leonid Pishchulin","Anton Milan","Juergen Gall","Bernt Schiele"],"abstract":"Human poses and motions are important cues for analysis of videos with people\nand there is strong evidence that representations based on body pose are highly\neffective for a variety of tasks such as activity recognition, content\nretrieval and social signal processing. In this work, we aim to further advance\nthe state of the art by establishing \"PoseTrack\", a new large-scale benchmark\nfor video-based human pose estimation and articulated tracking, and bringing\ntogether the community of researchers working on visual human analysis. The\nbenchmark encompasses three competition tracks focusing on i) single-frame\nmulti-person pose estimation, ii) multi-person pose estimation in videos, and\niii) multi-person articulated tracking. To facilitate the benchmark and\nchallenge we collect, annotate and release a new %large-scale benchmark dataset\nthat features videos with multiple people labeled with person tracks and\narticulated pose. A centralized evaluation server is provided to allow\nparticipants to evaluate on a held-out test set. We envision that the proposed\nbenchmark will stimulate productive research both by providing a large and\nrepresentative training dataset as well as providing a platform to objectively\nevaluate and compare the proposed methods. The benchmark is freely accessible\nat https://posetrack.net.","url_abs":"http://arxiv.org/abs/1710.10000v2","url_pdf":"http://arxiv.org/pdf/1710.10000v2.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":"posetrack-a-benchmark-for-human-pose","repo_url":"https://github.com/ken724049/action-recognition","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"posetrack-a-benchmark-for-human-pose","repo_url":"https://github.com/open-mmlab/mmpose","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"activity-recognition","task_name":"Activity Recognition"},{"task_slug":"multi-person-pose-estimation","task_name":"Multi-Person Pose Estimation"},{"task_slug":"pose-estimation","task_name":"Pose Estimation"},{"task_slug":"pose-tracking","task_name":"Pose Tracking"},{"task_slug":"retrieval","task_name":"Retrieval"}],"methods":[],"datasets_introduced":[{"slug":"posetrack","name":"PoseTrack","full_name":""}],"methods_introduced":[],"results":[{"leaderboard":"/sota/multi-person-pose-estimation-on-posetrack2017","task":"Multi-Person Pose Estimation","dataset":"PoseTrack2017","model":"PoseTrack","rank_in_archive_order":3,"of":3,"metrics":{"Mean mAP":"59.4"},"uses_additional_data":false},{"leaderboard":"/sota/pose-tracking-on-posetrack2017","task":"Pose Tracking","dataset":"PoseTrack2017","model":"PoseTrack","rank_in_archive_order":10,"of":10,"metrics":{"MOTA":"48.37","mAP":"59.22"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1710.10000","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}