Spatio-Temporal Attention-Based LSTM Networks for 3D Action Recognition and Detection

doi:10.1109/TIP.2018.2818328

CORC > 自动化研究所 > 中国科学院自动化研究所 > 模式识别国家重点实验室 > 视频内容安全团队

	Spatio-Temporal Attention-Based LSTM Networks for 3D Action Recognition and Detection
	Song, Sijie 1; Lan, Cuiling 2; Xing, Junliang4 ; Zeng, Wenjun 2,3; Liu, Jiaying 1
刊名	IEEE TRANSACTIONS ON IMAGE PROCESSING
	2018-07-01
卷号	27 期号:7 页码:3459-3471
关键词	Spatio Attention Temporal Attention Action Recognition Action Detection Skeleton Data
DOI	10.1109/TIP.2018.2818328
文献子类	Article
英文摘要	Human action analytics has attracted a lot of attention for decades in computer vision. It is important to extract discriminative spatio-temporal features to model the spatial and temporal evolutions of different actions. In this paper, we propose a spatial and temporal attention model to explore the spatial and temporal discriminative features for human action recognition and detection from skeleton data. We build our networks based on the recurrent neural networks with long short-term memory units. The learned model is capable of selectively focusing on discriminative joints of skeletons within each input frame and paying different levels of attention to the outputs of different frames. To ensure effective training of the network for action recognition, we propose a regularized cross-entropy loss to drive the learning process and develop a joint training strategy accordingly. Moreover, based on temporal attention, we develop a method to generate the action temporal proposals for action detection. We evaluate the proposed method on the SBU Kinect Interaction data set, the NTU RGB + D data set, and the PKU-MMD data set, respectively. Experiment results demonstrate the effectiveness of our proposed model on both action recognition and action detection.
WOS关键词	MOTION ; MODEL
WOS研究方向	Computer Science ; Engineering
语种	英语
WOS记录号	WOS:000430594300008
资助机构	National Natural Science Foundation of China(61772043 ; Microsoft Research Asia Fund(FY17-RES-THEME-013) ; CCF-Tencent Open Research Fund ; 61672519)
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/22007]
专题	自动化研究所_模式识别国家重点实验室_视频内容安全团队
作者单位	1.Peking Univ, Inst Comp Sci & Technol, Beijing 100080, Peoples R China 2.Microsoft Res Asia, Beijing 100080, Peoples R China 3.Microsoft Res Asia, Senior Leadership Team, Beijing 100080, Peoples R China 4.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100080, Peoples R China
推荐引用方式 GB/T 7714	Song, Sijie,Lan, Cuiling,Xing, Junliang,et al. Spatio-Temporal Attention-Based LSTM Networks for 3D Action Recognition and Detection[J]. IEEE TRANSACTIONS ON IMAGE PROCESSING,2018,27(7):3459-3471.
APA	Song, Sijie,Lan, Cuiling,Xing, Junliang,Zeng, Wenjun,&Liu, Jiaying.(2018).Spatio-Temporal Attention-Based LSTM Networks for 3D Action Recognition and Detection.IEEE TRANSACTIONS ON IMAGE PROCESSING,27(7),3459-3471.
MLA	Song, Sijie,et al."Spatio-Temporal Attention-Based LSTM Networks for 3D Action Recognition and Detection".IEEE TRANSACTIONS ON IMAGE PROCESSING 27.7(2018):3459-3471.