{"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/a-dual-source-approach-for-3d-human-pose","title":"A Dual-Source Approach for 3D Human Pose Estimation from a Single Image","arxiv_id":"1705.02883","date":"2017-05-08","proceeding":null,"authors":["Umar Iqbal","Andreas Doering","Hashim Yasin","Björn Krüger","Andreas Weber","Juergen Gall"],"abstract":"In this work we address the challenging problem of 3D human pose estimation\nfrom single images. Recent approaches learn deep neural networks to regress 3D\npose directly from images. One major challenge for such methods, however, is\nthe collection of training data. Specifically, collecting large amounts of\ntraining data containing unconstrained images annotated with accurate 3D poses\nis infeasible. We therefore propose to use two independent training sources.\nThe first source consists of accurate 3D motion capture data, and the second\nsource consists of unconstrained images with annotated 2D poses. To integrate\nboth sources, we propose a dual-source approach that combines 2D pose\nestimation with efficient 3D pose retrieval. To this end, we first convert the\nmotion capture data into a normalized 2D pose space, and separately learn a 2D\npose estimation model from the image data. During inference, we estimate the 2D\npose and efficiently retrieve the nearest 3D poses. We then jointly estimate a\nmapping from the 3D pose space to the image and reconstruct the 3D pose. We\nprovide a comprehensive evaluation of the proposed method and experimentally\ndemonstrate the effectiveness of our approach, even when the skeleton\nstructures of the two sources differ substantially.","url_abs":"http://arxiv.org/abs/1705.02883v2","url_pdf":"http://arxiv.org/pdf/1705.02883v2.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":"2d-pose-estimation","task_name":"2D Pose Estimation"},{"task_slug":"3d-human-pose-estimation","task_name":"3D Human Pose Estimation"},{"task_slug":"monocular-3d-human-pose-estimation","task_name":"Monocular 3D Human Pose Estimation"},{"task_slug":"pose-estimation","task_name":"Pose Estimation"},{"task_slug":"pose-retrieval","task_name":"Pose Retrieval"},{"task_slug":"retrieval","task_name":"Retrieval"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/monocular-3d-human-pose-estimation-on-human3","task":"Monocular 3D Human Pose Estimation","dataset":"Human3.6M","model":"Dual-source approach","rank_in_archive_order":39,"of":52,"metrics":{"Average MPJPE (mm)":"97.39","Frames Needed":"1","Need Ground Truth 2D Pose":"No","Use Video Sequence":"No"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1705.02883","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}