Skeleton-Based Mutual Action Recognition Using Interactive Skeleton Graph and Joint Attention
Paper
Paper/Presentation Title | Skeleton-Based Mutual Action Recognition Using Interactive Skeleton Graph and Joint Attention |
---|---|
Presentation Type | Paper |
Authors | Jia, Xiangze, Zhang, Ji, Wang, Zhen, Luo, Yonglong, Chen, Fulong and Yang, Gaoming |
Journal or Proceedings Title | Proceedings of 33rd International Conference on Database and Expert Systems Applications (DEXA 2022) |
Journal Citation | 13427, pp. 110-116 |
Number of Pages | 7 |
Year | 2022 |
Publisher | Springer |
Place of Publication | Switzerland |
ISBN | 9783031124259 |
9783031124266 | |
Digital Object Identifier (DOI) | https://doi.org/10.1007/978-3-031-12426-6_9 |
Web Address (URL) of Paper | https://link.springer.com/chapter/10.1007/978-3-031-12426-6_9 |
Web Address (URL) of Conference Proceedings | https://link.springer.com/book/10.1007/978-3-031-12426-6 |
Conference/Event | 33rd International Conference on Database and Expert Systems Applications (DEXA 2022) |
Event Details | 33rd International Conference on Database and Expert Systems Applications (DEXA 2022) Parent International Conference on Database and Expert Systems Applications Delivery In person Event Date 22 to end of 24 Aug 2022 Event Location Vienna, Austria |
Abstract | Skeleton-based action recognition relies on skeleton sequences to detect certain predetermined types of human actions. The existing related works are inadequate in mutual action recognition. We thus propose an innovative interactive skeleton graph to represent the skeleton data. In addition, because the GCN pays attention to the information about the edges in the skeleton graph which represent the interaction between joints, we propose a joint attention module that assists the model in paying attention to the pattern of vertices which represent the joints in the skeleton graph. We validate our model on the NTU RGB-D datasets, and the experimental results demonstrate the superiority of our model against other baseline methods in terms of recognition effectiveness in understanding mutual actions. |
Keywords | Interactive skeleton graph ; Action recognition; Joint attention |
ANZSRC Field of Research 2020 | 460999. Information systems not elsewhere classified |
Public Notes | Files associated with this item cannot be displayed due to copyright restrictions. |
Series | Lecture Notes in Computer Science |
Byline Affiliations | Nanjing University of Aeronautics and Astronautics, China |
University of Southern Queensland | |
Zhejiang Lab, China | |
Anhui Normal University, China |
https://research.usq.edu.au/item/z58yz/skeleton-based-mutual-action-recognition-using-interactive-skeleton-graph-and-joint-attention
67
total views2
total downloads1
views this month0
downloads this month