{"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/learning-pose-grammar-to-encode-human-body","title":"Learning Pose Grammar to Encode Human Body Configuration for 3D Pose Estimation","arxiv_id":"1710.06513","date":"2017-10-17","proceeding":null,"authors":["Hao-Shu Fang","Yuanlu Xu","Wenguan Wang","Xiaobai Liu","Song-Chun Zhu"],"abstract":"In this paper, we propose a pose grammar to tackle the problem of 3D human\npose estimation. Our model directly takes 2D pose as input and learns a\ngeneralized 2D-3D mapping function. The proposed model consists of a base\nnetwork which efficiently captures pose-aligned features and a hierarchy of\nBi-directional RNNs (BRNN) on the top to explicitly incorporate a set of\nknowledge regarding human body configuration (i.e., kinematics, symmetry, motor\ncoordination). The proposed model thus enforces high-level constraints over\nhuman poses. In learning, we develop a pose sample simulator to augment\ntraining samples in virtual camera views, which further improves our model\ngeneralizability. We validate our method on public 3D human pose benchmarks and\npropose a new evaluation protocol working on cross-view setting to verify the\ngeneralization capability of different methods. We empirically observe that\nmost state-of-the-art methods encounter difficulty under such setting while our\nmethod can well handle such challenges.","url_abs":"http://arxiv.org/abs/1710.06513v6","url_pdf":"http://arxiv.org/pdf/1710.06513v6.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":"3d-absolute-human-pose-estimation","task_name":"3D Absolute Human Pose Estimation"},{"task_slug":"3d-human-pose-estimation","task_name":"3D Human Pose Estimation"},{"task_slug":"3d-pose-estimation","task_name":"3D Pose Estimation"},{"task_slug":"pose-estimation","task_name":"Pose Estimation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/3d-absolute-human-pose-estimation-on-human36m","task":"3D Absolute Human Pose Estimation","dataset":"Human3.6M","model":"Pose Grammar","rank_in_archive_order":3,"of":4,"metrics":{"Average MPJPE (mm)":"60.4"},"uses_additional_data":false},{"leaderboard":"/sota/3d-human-pose-estimation-on-humaneva-i","task":"3D Human Pose Estimation","dataset":"HumanEva-I","model":"Pose Grammar","rank_in_archive_order":16,"of":31,"metrics":{"Mean Reconstruction Error (mm)":"22.9"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1710.06513","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}