{"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/human-pose-estimation-with-spatial-contextual","title":"Human Pose Estimation with Spatial Contextual Information","arxiv_id":"1901.01760","date":"2019-01-07","proceeding":null,"authors":["Hong Zhang","Hao Ouyang","Shu Liu","Xiaojuan Qi","Xiaoyong Shen","Ruigang Yang","Jiaya Jia"],"abstract":"We explore the importance of spatial contextual information in human pose\nestimation. Most state-of-the-art pose networks are trained in a multi-stage\nmanner and produce several auxiliary predictions for deep supervision. With\nthis principle, we present two conceptually simple and yet computational\nefficient modules, namely Cascade Prediction Fusion (CPF) and Pose Graph Neural\nNetwork (PGNN), to exploit underlying contextual information. Cascade\nprediction fusion accumulates prediction maps from previous stages to extract\ninformative signals. The resulting maps also function as a prior to guide\nprediction at following stages. To promote spatial correlation among joints,\nour PGNN learns a structured representation of human pose as a graph. Direct\nmessage passing between different joints is enabled and spatial relation is\ncaptured. These two modules require very limited computational complexity.\nExperimental results demonstrate that our method consistently outperforms\nprevious methods on MPII and LSP benchmark.","url_abs":"http://arxiv.org/abs/1901.01760v1","url_pdf":"http://arxiv.org/pdf/1901.01760v1.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":"graph-neural-network","task_name":"Graph Neural Network"},{"task_slug":"pose-estimation","task_name":"Pose Estimation"},{"task_slug":"prediction","task_name":"Prediction"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/pose-estimation-on-mpii-human-pose","task":"Pose Estimation","dataset":"MPII Human Pose","model":"Spatial Context","rank_in_archive_order":10,"of":46,"metrics":{"PCKh-0.5":"92.5"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1901.01760","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}