文献类型：期刊文献

英文题名：PGCN-TCA: Pseudo Graph Convolutional Network With Temporal and Channel-Wise Attention for Skeleton-Based Action Recognition

作者：Yang, Hongye[1,2];Gu, Yuzhang[1,2];Zhu, Jianchao[3];Hu, Keli[4];Zhang, Xiaolin[1,2,5]

机构：[1]Chinese Acad Sci, Shanghai Inst Microsyst & Informat Technol, Biovis Syst Lab, State Key Lab Transducer Technol, Shanghai 200050, Peoples R China;[2]Univ Chinese Acad Sci, Sch Elect Elect & Commun Engn, Beijing 100049, Peoples R China;[3]East China Normal Univ, Sch Comp Sci & Software Engn, Shanghai 200062, Peoples R China;[4]Shaoxing Univ, Dept Comp Sci & Engn, Shaoxing 312000, Peoples R China;[5]ShanghaiTech Univ, Sch Informat Sci & Technol, Shanghai 201210, Peoples R China

年份：2020

卷号：8

起止页码：10040

外文期刊名：IEEE ACCESS

收录：SCI-EXPANDED(收录号：WOS:000549569100005)、、EI(收录号：20200508102710)、Scopus(收录号：2-s2.0-85078488517)、WOS

基金：This work was supported in part by the National Key Research and Development Program of China under Grant 2017YFC0805500, in part by the Shanghai Municipal Science and Technology Major Project (ZHANGJIANG LAB) under Grant 2018SHZDZX01, and in part by the National Natural Science Foundation of China under Grant 61603258.

语种：英文

外文关键词：Computer vision; skeleton-based action recognition; temporal and channel-wise attention

外文摘要：Skeleton-based human action recognition has become an active research area in recent years. The key to this task is to fully explore both spatial and temporal features. Recently, GCN-based methods modeling the human body skeletons as spatial-temporal graphs, have achieved remarkable performances. However, most GCN-based methods use a fixed adjacency matrix defined by the dataset, which can only capture the structural information provided by joints directly connected through bones and ignore the dependencies between distant joints that are not connected. In addition, such a fixed adjacency matrix used in all layers leads to the network failing to extract multi-level semantic features. In this paper we propose a pseudo graph convolutional network with temporal and channel-wise attention (PGCN-TCA) to solve this problem. The fixed normalized adjacent matrix is substituted with a learnable matrix. In this way, the matrix can learn the dependencies between connected joints and joints that are not physically connected. At the same time, learnable matrices in different layers can help the network capture multi-level features in spatial domain. Moreover, Since frames and input channels that contain outstanding characteristics play significant roles in distinguishing the action from others, we propose a mixed temporal and channel-wise attention. Our method achieves comparable performances to state-of-the-art methods on NTU-RGB & x002B;D and HDM05 datasets.