Improving skeleton-based action recognition with interactive object information

• Published in: International journal of multimedia information retrieval Vol. 14; no. 1; p. 3
• Main Authors: Wen, Hao, Lu, Ziqian, Shen, Fengli, Lu, Zhe-Ming, Cui, Jialin
• Format: Journal Article
• Language: English
• Published: London Springer London 01.03.2025; Springer Nature B.V
• Subjects: Activity recognition; Artificial neural networks; Computer Science; Data augmentation; Data Mining and Knowledge Discovery; Database Management; Datasets; Digitization; Factories; Graphs; Image Processing and Computer Vision; Information Storage and Retrieval; Information Systems Applications (incl.Internet); Methods; Moving object recognition; Multimedia Information Systems; Nodes; Performance evaluation; Regular Paper; Action Recognition; Human-Object Interaction; Smart Factory; Graph Convolutional Networks
• ISSN: 2192-6611; 2192-662X
• DOI: 10.1007/s13735-024-00351-7

Human skeleton information is important in skeleton-based action recognition, which provides a simple and efficient way to describe human pose. However, existing skeleton-based methods focus more on the skeleton, ignoring the objects interacting with humans, resulting in poor performance in recognizing actions that involve object interactions. We propose a new action recognition framework introducing object nodes to supplement absent interactive object information. We also propose Spatial Temporal Variable Graph Convolutional Networks (ST-VGCN) to effectively model the Variable Graph (VG) containing object nodes. Specifically, in order to validate the role of interactive object information, by leveraging a simple self-training approach, we establish a new dataset, JXGC 24, and an extended dataset, NTU RGB+D+Object 60, including more than 2 million additional object nodes. At the same time, we designe the Variable Graph construction method to accommodate a variable number of nodes for graph structure. Additionally, we are the first to explore the overfitting issue introduced by incorporating additional object information, and we propose a VG-based data augmentation method to address this issue, called Random Node Attack. Finally, regarding the network structure, we introduce two fusion modules, CAF and WNPool, along with a novel Node Balance Loss, to enhance the comprehensive performance by effectively fusing and balancing skeleton and object node information. Our method surpasses the previous state-of-the-art on multiple skeleton-based action recognition benchmarks. The accuracy of our method on NTU RGB+D 60 cross-subject split is 96.7%, and on cross-view split, it is 99.2%. The project page: https://github.com/moonlight52137/ST-VGCN .