CASIA OpenIR  > 智能感知与计算研究中心
Learning robust representations using recurrent neural networks for skeleton based action classification and detection
Wang Hongsong(王洪松)1,2,3,4; Liang Wang(王亮)1,2,3,4
2017
Conference NameIEEE International Conference on Multimedia & Expo(ICME) Workshops
Conference Date10-14 July 2017
Conference PlaceHong Kong
Abstract
Recently, skeleton based action recognition gains more popularity due to affordable depth sensors and real-time skeleton estimation algorithms. Previous Recurrent Neural Networks (RNN) based approaches focus on modeling spatial configuration of skeletons and temporal evolution of body joints. There are certain intrinsic characteristics of the skeleton based actions. For example, the starting point may be varied, an action can be observed at arbitrary viewpoints and the skeletons are noisy. To this end, we present a novel end-to-end architecture based on RNN to learn robust representations from raw skeletons. The architecture includes three new layers, i.e., starting point transformation layer, viewpoint transformation layer and spatial dropout layer, which address the corresponding three problems, respectively. We apply the proposed method to two different tasks: action classification and detection. Experiments on two large-scale datasets (NTU RGB+D and PKU-MMD) show the superiority of our model. Specially, for action detection, our results are more than 33.4% higher the previous results.
Document Type会议论文
Identifierhttp://ir.ia.ac.cn/handle/173211/19625
Collection智能感知与计算研究中心
Affiliation1.Center for Research on Intelligent Perception and Computing (CRIPAC)
2.National Laboratory of Pattern Recognition (NLPR)
3.Institute of Automation, Chinese Academy of Sciences (CASIA)
4.University of Chinese Academy of Sciences (UCAS)
Recommended Citation
GB/T 7714
Wang Hongsong,Liang Wang. Learning robust representations using recurrent neural networks for skeleton based action classification and detection[C],2017.
Files in This Item: Download All
File Name/Size DocType Version Access License
LEARNING ROBUST REPR(473KB)会议论文 开放获取CC BY-NC-SAView Download
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[Wang Hongsong(王洪松)]'s Articles
[Liang Wang(王亮)]'s Articles
Baidu academic
Similar articles in Baidu academic
[Wang Hongsong(王洪松)]'s Articles
[Liang Wang(王亮)]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[Wang Hongsong(王洪松)]'s Articles
[Liang Wang(王亮)]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: LEARNING ROBUST REPRESENTATIONS USING RECURRENT NEURAL NETWORKS FOR SKELETON BASED ACTION CLASSIFICATION A ND DETECTION.pdf
Format: Adobe PDF
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.