高技术通讯2025，Vol.35Issue(1)：37-46,10.DOI:10.3772/j.issn.1002-0470.2025.01.004
基于多尺度图卷积网络的骨架行为识别方法
Multi-scale graph convolution network for skeleton-based action recognition
摘要
Abstract
In view of the lack of modeling of the relationship between remote nodes in existing skeleton-based action rec-ognition methods,which leads to low recognition accuracy and poor generalization ability,a new action recognition method based on multi-scale graph convolutional network(MS-GCN)is proposed.Firstly,the convolution of ex-tended graphs is constructed by multi-hop adjacency matrix,and the convolution of multi-scale spatial graphs is constructed by combining the convolution of expanded graphs with different hop numbers.Secondly,spatial channel attention(SCA)is proposed to stimulate spatial sensitive channels to further enhance spatial features.Finally,the uniform sampling data enhancement method is used to generate diversified training samples to enhance the robust-ness and generalization ability of the model.The accuracy of the proposed method is 97.24%(X-View),90.43%(X-Set)and 96.34%on the data sets NTU-RGB+D 60,NTU-RGB+D 120 and Northwestern-UCLA,respective-ly,which fully verifies the effectiveness of the proposed method.关键词
骨架行为识别/多尺度/图卷积/空间通道注意力Key words
skeleton-based action recognition/multi-scale/graph convolution/spatial channel attention引用本文复制引用
安文志,冯宇平,李云文,赵军,董金宇..基于多尺度图卷积网络的骨架行为识别方法[J].高技术通讯,2025,35(1):37-46,10.基金项目
国家自然科学基金(61971253)和国家级大学生创新创业训练项目(202310426296,202310426356)资助. （61971253）
