DSANet: A lightweight hybrid network for human action recognition in virtual sports

• Published in: Computer animation and virtual worlds Vol. 35; no. 3
• Main Authors: Xiao, Zhiyong, Yu, Feng, Liu, Li, Peng, Tao, Hu, Xinrong, Jiang, Minghua
• Format: Journal Article
• Language: English
• Published: Chichester Wiley Subscription Services, Inc 01.05.2024
• Subjects: Accuracy; activity recognition; attention mechanism; Complexity; Feature extraction; Human activity recognition; Lightweight; lightweight network; Modules; multiscale feature; Networks; Parameters; Spatial data; virtual sports; Weight reduction
• ISSN: 1546-4261; 1546-427X
• DOI: 10.1002/cav.2274

Human activity recognition (HAR) has significant potential in virtual sports applications. However, current HAR networks often prioritize high accuracy at the expense of practical application requirements, resulting in networks with large parameter counts and computational complexity. This can pose challenges for real‐time and efficient recognition. This paper proposes a hybrid lightweight DSANet network designed to address the challenges of real‐time performance and algorithmic complexity. The network utilizes a multi‐scale depthwise separable convolutional (Multi‐scale DWCNN) module to extract spatial information and a multi‐layer Gated Recurrent Unit (Multi‐layer GRU) module for temporal feature extraction. It also incorporates an improved channel‐space attention module called RCSFA to enhance feature extraction capability. By leveraging channel, spatial, and temporal information, the network achieves a low number of parameters with high accuracy. Experimental evaluations on UCIHAR, WISDM, and PAMAP2 datasets demonstrate that the network not only reduces parameter counts but also achieves accuracy rates of 97.55%, 98.99%, and 98.67%, respectively, compared to state‐of‐the‐art networks. This research provides valuable insights for the virtual sports field and presents a novel network for real‐time activity recognition deployment in embedded devices. A hybrid lightweight network called DSANet, designed to address the challenges of real‐time performance and algorithmic complexity in virtual sports. DSANet incorporates a multi‐scale depthwise separable convolutional (Multi‐scale DWCNN) module for spatial feature extraction and a multi‐layer Gated Recurrent Unit (Multi‐layer GRU) module for temporal feature extraction. Additionally, it incorporates an improved channel‐space attention module named RCSFA to enhance feature extraction capability. The network achieves high accuracy with a low number of parameters and outperforms state‐of‐the‐art networks on multiple datasets, as demonstrated by experimental evaluations. This paper provides valuable insights for the virtual sports field and presents a novel network suitable for real‐time activity recognition deployment on embedded devices.