{"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/view-adaptive-recurrent-neural-networks-for","title":"View Adaptive Recurrent Neural Networks for High Performance Human Action Recognition from Skeleton Data","arxiv_id":"1703.08274","date":"2017-03-24","proceeding":"ICCV 2017 10","authors":["Pengfei Zhang","Cuiling Lan","Junliang Xing","Wen-Jun Zeng","Jianru Xue","Nanning Zheng"],"abstract":"Skeleton-based human action recognition has recently attracted increasing\nattention due to the popularity of 3D skeleton data. One main challenge lies in\nthe large view variations in captured human actions. We propose a novel view\nadaptation scheme to automatically regulate observation viewpoints during the\noccurrence of an action. Rather than re-positioning the skeletons based on a\nhuman defined prior criterion, we design a view adaptive recurrent neural\nnetwork (RNN) with LSTM architecture, which enables the network itself to adapt\nto the most suitable observation viewpoints from end to end. Extensive\nexperiment analyses show that the proposed view adaptive RNN model strives to\n(1) transform the skeletons of various views to much more consistent viewpoints\nand (2) maintain the continuity of the action rather than transforming every\nframe to the same position with the same body orientation. Our model achieves\nsignificant improvement over the state-of-the-art approaches on three benchmark\ndatasets.","url_abs":"http://arxiv.org/abs/1703.08274v2","url_pdf":"http://arxiv.org/pdf/1703.08274v2.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":"view-adaptive-recurrent-neural-networks-for","repo_url":"https://github.com/microsoft/View-Adaptive-Neural-Networks-for-Skeleton-based-Human-Action-Recognition","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"action-recognition-in-videos","task_name":"Action Recognition"},{"task_slug":"skeleton-based-action-recognition","task_name":"Skeleton Based Action Recognition"},{"task_slug":"action-recognition","task_name":"Temporal Action Localization"}],"methods":[{"method_slug":"lstm","method_name":"LSTM"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"tanh-activation","method_name":"Tanh Activation"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/skeleton-based-action-recognition-on-ntu-rgbd","task":"Skeleton Based Action Recognition","dataset":"NTU RGB+D","model":"VA-LSTM","rank_in_archive_order":117,"of":135,"metrics":{"Accuracy (CS)":"79.2","Accuracy (CV)":"87.6"},"uses_additional_data":false},{"leaderboard":"/sota/skeleton-based-action-recognition-on-sysu-3d","task":"Skeleton Based Action Recognition","dataset":"SYSU 3D","model":"VA-LSTM","rank_in_archive_order":6,"of":9,"metrics":{"Accuracy":"77.5%"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1703.08274","atlas_url":"https://app.syntology.ai/?focus=1703.08274","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}