{"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/bodynet-volumetric-inference-of-3d-human-body","title":"BodyNet: Volumetric Inference of 3D Human Body Shapes","arxiv_id":"1804.04875","date":"2018-04-13","proceeding":"ECCV 2018 9","authors":["Gül Varol","Duygu Ceylan","Bryan Russell","Jimei Yang","Ersin Yumer","Ivan Laptev","Cordelia Schmid"],"abstract":"Human shape estimation is an important task for video editing, animation and\nfashion industry. Predicting 3D human body shape from natural images, however,\nis highly challenging due to factors such as variation in human bodies,\nclothing and viewpoint. Prior methods addressing this problem typically attempt\nto fit parametric body models with certain priors on pose and shape. In this\nwork we argue for an alternative representation and propose BodyNet, a neural\nnetwork for direct inference of volumetric body shape from a single image.\nBodyNet is an end-to-end trainable network that benefits from (i) a volumetric\n3D loss, (ii) a multi-view re-projection loss, and (iii) intermediate\nsupervision of 2D pose, 2D body part segmentation, and 3D pose. Each of them\nresults in performance improvement as demonstrated by our experiments. To\nevaluate the method, we fit the SMPL model to our network output and show\nstate-of-the-art results on the SURREAL and Unite the People datasets,\noutperforming recent approaches. Besides achieving state-of-the-art\nperformance, our method also enables volumetric body-part segmentation.","url_abs":"http://arxiv.org/abs/1804.04875v3","url_pdf":"http://arxiv.org/pdf/1804.04875v3.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":"bodynet-volumetric-inference-of-3d-human-body","repo_url":"https://github.com/LONG-9621/body_net","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"bodynet-volumetric-inference-of-3d-human-body","repo_url":"https://github.com/gulvarol/bodynet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"3d-human-pose-estimation","task_name":"3D Human Pose Estimation"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"video-editing","task_name":"Video Editing"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/3d-human-pose-estimation-on-surreal-1","task":"3D Human Pose Estimation","dataset":"Surreal","model":"BodyNet","rank_in_archive_order":3,"of":6,"metrics":{"MPJPE":"49.1"},"uses_additional_data":true}],"syntology":{"syntology_url":"https://syntology.ai/paper/1804.04875","atlas_url":"https://app.syntology.ai/?focus=1804.04875","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}