{"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/generating-multiple-hypotheses-for-3d-human","title":"Generating Multiple Hypotheses for 3D Human Pose Estimation with Mixture Density Network","arxiv_id":"1904.05547","date":"2019-04-11","proceeding":"CVPR 2019 6","authors":["Chen Li","Gim Hee Lee"],"abstract":"3D human pose estimation from a monocular image or 2D joints is an ill-posed\nproblem because of depth ambiguity and occluded joints. We argue that 3D human\npose estimation from a monocular input is an inverse problem where multiple\nfeasible solutions can exist. In this paper, we propose a novel approach to\ngenerate multiple feasible hypotheses of the 3D pose from 2D joints.In contrast\nto existing deep learning approaches which minimize a mean square error based\non an unimodal Gaussian distribution, our method is able to generate multiple\nfeasible hypotheses of 3D pose based on a multimodal mixture density networks.\nOur experiments show that the 3D poses estimated by our approach from an input\nof 2D joints are consistent in 2D reprojections, which supports our argument\nthat multiple solutions exist for the 2D-to-3D inverse problem. Furthermore, we\nshow state-of-the-art performance on the Human3.6M dataset in both best\nhypothesis and multi-view settings, and we demonstrate the generalization\ncapacity of our model by testing on the MPII and MPI-INF-3DHP datasets. Our\ncode is available at the project website.","url_abs":"http://arxiv.org/abs/1904.05547v1","url_pdf":"http://arxiv.org/pdf/1904.05547v1.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":"generating-multiple-hypotheses-for-3d-human","repo_url":"https://github.com/chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"generating-multiple-hypotheses-for-3d-human","repo_url":"https://github.com/vnmr/JointVideoPose3D","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"gone","observed_at":"2026-09-18","how":"tree_404+repo_404"}}],"tasks":[{"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":"multi-hypotheses-3d-human-pose-estimation","task_name":"Multi-Hypotheses 3D Human Pose Estimation"},{"task_slug":"pose-estimation","task_name":"Pose Estimation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/3d-human-pose-estimation-on-human36m","task":"3D Human Pose Estimation","dataset":"Human3.6M","model":"MDN (Multi-View)","rank_in_archive_order":61,"of":88,"metrics":{"Average MPJPE (mm)":"49.6","Multi-View or Monocular":"Multi-View","Using 2D ground-truth joints":"No"},"uses_additional_data":false},{"leaderboard":"/sota/3d-human-pose-estimation-on-human36m","task":"3D Human Pose Estimation","dataset":"Human3.6M","model":"MDN","rank_in_archive_order":77,"of":88,"metrics":{"Average MPJPE (mm)":"52.7","Multi-View or Monocular":"Monocular","PA-MPJPE":"42.6","Using 2D ground-truth joints":"No"},"uses_additional_data":false},{"leaderboard":"/sota/3d-human-pose-estimation-on-mpi-inf-3dhp","task":"3D Human Pose Estimation","dataset":"MPI-INF-3DHP","model":"MDM","rank_in_archive_order":107,"of":108,"metrics":{"PCK":"67.9"},"uses_additional_data":false},{"leaderboard":"/sota/monocular-3d-human-pose-estimation-on-human3","task":"Monocular 3D Human Pose Estimation","dataset":"Human3.6M","model":"Multimodal Mixture Density Networks","rank_in_archive_order":27,"of":52,"metrics":{"Average MPJPE (mm)":"52.7","Frames Needed":"1","Need Ground Truth 2D Pose":"No","Use Video Sequence":"No"},"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-MDN (by 3D Multi-bodies)","rank_in_archive_order":5,"of":10,"metrics":{"Best-Hypothesis MPJPE (n = 25)":"91.5","Best-Hypothesis PMPJPE (n = 25)":"69.5","H36M PMPJPE (n = 1)":"44.8","H36M PMPJPE (n = 25)":"42.7","Most-Likely Hypothesis PMPJPE (n = 1)":"74.7"},"uses_additional_data":false},{"leaderboard":"/sota/multi-hypotheses-3d-human-pose-estimation-on","task":"Multi-Hypotheses 3D Human Pose Estimation","dataset":"Human3.6M","model":"MDN","rank_in_archive_order":9,"of":12,"metrics":{"Average MPJPE (mm)":"52.7","Average PMPJPE (mm)":"42.6"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1904.05547","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1904.05547"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/vnmr/JointVideoPose3D","reach":{"status":"gone","observed_at":"2026-09-18","how":"tree_404+repo_404"}}],"summary":{"ran_draft_wrong":1,"unverified":10},"by_repo_kind":{"official":{"samples":11,"ran":1,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"f5eee6f4a7a2d3ce","entry":"get_action_subset","repo":"chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network","repo_kind":"official","path":"src/predict_3dpose_mdm.py","file_url":"https://github.com/chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network/blob/HEAD/src/predict_3dpose_mdm.py","link_basis":"plan_row","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"f5eee6f4a7a2d3ce"}},{"code_sha256_prefix":"de6eedce004eb352","entry":"Dirichlet_loss","repo":"chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network","repo_kind":"official","path":"src/mix_den_model.py","file_url":"https://github.com/chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network/blob/HEAD/src/mix_den_model.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"de6eedce004eb352"}},{"code_sha256_prefix":"6ce042354cb0a857","entry":"camera_to_world_frame","repo":"chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network","repo_kind":"official","path":"src/cameras.py","file_url":"https://github.com/chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network/blob/HEAD/src/cameras.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"6ce042354cb0a857"}},{"code_sha256_prefix":"d15a574f20fa5bea","entry":"compute_similarity_transform","repo":"chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network","repo_kind":"official","path":"src/procrustes.py","file_url":"https://github.com/chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network/blob/HEAD/src/procrustes.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"d15a574f20fa5bea"}},{"code_sha256_prefix":"a8756021030e5968","entry":"kaiming","repo":"chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network","repo_kind":"official","path":"src/mix_den_model.py","file_url":"https://github.com/chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network/blob/HEAD/src/mix_den_model.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"a8756021030e5968"}},{"code_sha256_prefix":"eba040e815a85e27","entry":"load_data","repo":"chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network","repo_kind":"official","path":"src/data_utils.py","file_url":"https://github.com/chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network/blob/HEAD/src/data_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"eba040e815a85e27"}},{"code_sha256_prefix":"32d5f4c81b3b566f","entry":"load_stacked_hourglass","repo":"chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network","repo_kind":"official","path":"src/data_utils.py","file_url":"https://github.com/chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network/blob/HEAD/src/data_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"32d5f4c81b3b566f"}},{"code_sha256_prefix":"a866706041e55d76","entry":"mean_log_Gaussian_like","repo":"chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network","repo_kind":"official","path":"src/mix_den_model.py","file_url":"https://github.com/chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network/blob/HEAD/src/mix_den_model.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"a866706041e55d76"}},{"code_sha256_prefix":"9088d315ea27a453","entry":"normalization_stats","repo":"chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network","repo_kind":"official","path":"src/data_utils.py","file_url":"https://github.com/chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network/blob/HEAD/src/data_utils.py","link_basis":"plan_row","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"9088d315ea27a453"}},{"code_sha256_prefix":"b4846342f739f10e","entry":"project_point_radial","repo":"chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network","repo_kind":"official","path":"src/cameras.py","file_url":"https://github.com/chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network/blob/HEAD/src/cameras.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"b4846342f739f10e"}},{"code_sha256_prefix":"c30df769883845b6","entry":"world_to_camera_frame","repo":"chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network","repo_kind":"official","path":"src/cameras.py","file_url":"https://github.com/chaneyddtt/Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network/blob/HEAD/src/cameras.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"c30df769883845b6"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}