Skeleton Action Recognition Based on Multi-Stream Spatial Attention Graph Convolutional SRU Network

Action recognition with skeleton data has attracted more attention. In order to solve the problems of low reasoning speed and single data mode of most algorithms, a lightweight and efficient method is proposed. The network em⁃ beds the graph convolution operator in the simple recurrent unit(SRU) to construct the graph convolutional SRU(GC-SRU), which can capture the spatial-temporal information of data. Meanwhile, to enhance the distinction between nodes, spatial at⁃ tention network and multi-stream data fusion are used to expand GC-SRU into multi-stream spatial attention graph convolu⁃ tional SRU(MSAGC-SRU). Finally, the proposed method is evaluated on two public datasets. Experimental results show that the classification accuracy of our method on Northwestern-UCLA reaches 93.1% and the FLOPs of the model is 4.4 G. The accuracy on NTU RGB+D reaches 92.7% and 87.3% under the CV and CS evaluation protocols, respectively, and the FLOPs of the model is 21.3 G. The proposed model has achieved good trade-off between computational efficiency and clas⁃ sification accuracy. © 2022 Chinese Institute of Electronics. All rights reserved.