{"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-human-pose-estimation-in-the-wild-by","title":"3D Human Pose Estimation in the Wild by Adversarial Learning","arxiv_id":"1803.09722","date":"2018-03-26","proceeding":"CVPR 2018 6","authors":["Wei Yang","Wanli Ouyang","Xiaolong Wang","Jimmy Ren","Hongsheng Li","Xiaogang Wang"],"abstract":"Recently, remarkable advances have been achieved in 3D human pose estimation\nfrom monocular images because of the powerful Deep Convolutional Neural\nNetworks (DCNNs). Despite their success on large-scale datasets collected in\nthe constrained lab environment, it is difficult to obtain the 3D pose\nannotations for in-the-wild images. Therefore, 3D human pose estimation in the\nwild is still a challenge. In this paper, we propose an adversarial learning\nframework, which distills the 3D human pose structures learned from the fully\nannotated dataset to in-the-wild images with only 2D pose annotations. Instead\nof defining hard-coded rules to constrain the pose estimation results, we\ndesign a novel multi-source discriminator to distinguish the predicted 3D poses\nfrom the ground-truth, which helps to enforce the pose estimator to generate\nanthropometrically valid poses even with images in the wild. We also observe\nthat a carefully designed information source for the discriminator is essential\nto boost the performance. Thus, we design a geometric descriptor, which\ncomputes the pairwise relative locations and distances between body joints, as\na new information source for the discriminator. The efficacy of our adversarial\nlearning framework with the new geometric descriptor has been demonstrated\nthrough extensive experiments on widely used public benchmarks. Our approach\nsignificantly improves the performance compared with previous state-of-the-art\napproaches.","url_abs":"http://arxiv.org/abs/1803.09722v2","url_pdf":"http://arxiv.org/pdf/1803.09722v2.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-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":null,"task_name":"valid"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/3d-human-pose-estimation-on-mpi-inf-3dhp","task":"3D Human Pose Estimation","dataset":"MPI-INF-3DHP","model":"Adversarial Learning","rank_in_archive_order":103,"of":108,"metrics":{"AUC":"32.0","PCK":"69.0"},"uses_additional_data":false},{"leaderboard":"/sota/monocular-3d-human-pose-estimation-on-human3","task":"Monocular 3D Human Pose Estimation","dataset":"Human3.6M","model":"Adversarial Learning","rank_in_archive_order":45,"of":52,"metrics":{"Frames Needed":"1","Need Ground Truth 2D Pose":"No","Use Video Sequence":"No"},"uses_additional_data":false},{"leaderboard":"/sota/pose-estimation-on-mpii-single-person","task":"Pose Estimation","dataset":"MPII Single Person","model":"Adversarial Learning","rank_in_archive_order":5,"of":5,"metrics":{"PCKh@0.5":"88.6"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1803.09722","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}