{"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/3d-multi-bodies-fitting-sets-of-plausible-3d","title":"3D Multi-bodies: Fitting Sets of Plausible 3D Human Models to Ambiguous Image Data","arxiv_id":"2011.00980","date":"2020-11-02","proceeding":"NeurIPS 2020 12","authors":["Benjamin Biggs","Sébastien Ehrhadt","Hanbyul Joo","Benjamin Graham","Andrea Vedaldi","David Novotny"],"abstract":"We consider the problem of obtaining dense 3D reconstructions of humans from single and partially occluded views. In such cases, the visual evidence is usually insufficient to identify a 3D reconstruction uniquely, so we aim at recovering several plausible reconstructions compatible with the input data. We suggest that ambiguities can be modelled more effectively by parametrizing the possible body shapes and poses via a suitable 3D model, such as SMPL for humans. We propose to learn a multi-hypothesis neural network regressor using a best-of-M loss, where each of the M hypotheses is constrained to lie on a manifold of plausible human poses by means of a generative model. We show that our method outperforms alternative approaches in ambiguous pose recovery on standard benchmarks for 3D humans, and in heavily occluded versions of these benchmarks.","url_abs":"https://arxiv.org/abs/2011.00980v1","url_pdf":"https://arxiv.org/pdf/2011.00980v1.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-reconstruction","task_name":"3D Reconstruction"},{"task_slug":"multi-hypotheses-3d-human-pose-estimation","task_name":"Multi-Hypotheses 3D Human Pose Estimation"}],"methods":[],"datasets_introduced":[{"slug":"ah36m","name":"AH36M","full_name":"Ambiguous Human3.6M"}],"methods_introduced":[],"results":[{"leaderboard":"/sota/multi-hypotheses-3d-human-pose-estimation-on-2","task":"Multi-Hypotheses 3D Human Pose Estimation","dataset":"AH36M","model":"3D Multi-bodies","rank_in_archive_order":3,"of":10,"metrics":{"Best-Hypothesis MPJPE (n = 25)":"90.0","Best-Hypothesis PMPJPE (n = 25)":"64.2","H36M PMPJPE (n = 1)":"41.6","H36M PMPJPE (n = 25)":"42.2","Most-Likely Hypothesis PMPJPE (n = 1)":"67.8"},"uses_additional_data":false},{"leaderboard":"/sota/multi-hypotheses-3d-human-pose-estimation-on-2","task":"Multi-Hypotheses 3D Human Pose Estimation","dataset":"AH36M","model":"SMPL-CVAE","rank_in_archive_order":6,"of":10,"metrics":{"Best-Hypothesis MPJPE (n = 25)":"109.7","Best-Hypothesis PMPJPE (n = 25)":"75.1","H36M PMPJPE (n = 1)":"46.7","H36M PMPJPE (n = 25)":"46.2","Most-Likely Hypothesis PMPJPE (n = 1)":"76.5"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2011.00980","atlas_url":"https://app.syntology.ai/?focus=2011.00980","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}