In order to solve the problems of low accuracy and poor robustness of traffic police dynamic gesture recognition, a ST-GCN based traffic police dynamic gesture recognition method is designed in this paper. The skeletal features of traffic police were extracted from space and time dimensions. By updating the graph attention matrix, the skeletal connectivity structure of traffic police was optimized to highlight the effective spatial features. Increase the time attention mechanism and strengthen the characteristics of the core movements in the learning gestures. The design model compares 8 types of gestures of traffic police in different scenarios, and the results show that: The ST-GCN traffic police gesture recognition network, which integrates spatio-temporal attention mechanism, achieves good recognition effect, with an average recognition accuracy of 88.82%, which is 5.79% higher than that of classical ST-GCN, which verifies the good performance of the proposed algorithm in traffic police dynamic gesture recognition.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Traffic Police Dynamic Gesture Recognition Based on Spatiotemporal Attention ST-GCN


    Additional title:

    Lect. Notes Electrical Eng.


    Contributors:
    Jia, Yingmin (editor) / Zhang, Weicun (editor) / Fu, Yongling (editor) / Wang, Jiqiang (editor) / Wu, Xiru (author) / Zhao, Yu (author) / Chen, Qi (author)

    Conference:

    Chinese Intelligent Systems Conference ; 2023 ; Ningbo, China October 14, 2023 - October 15, 2023



    Publication date :

    2023-10-08


    Size :

    22 pages





    Type of media :

    Article/Chapter (Book)


    Type of material :

    Electronic Resource


    Language :

    English





    Traffic Police Gesture Recognition Based on Openpose and GRU

    Wang, Dengwen / Wang, Wangmeng / Chen, Yanbing et al. | IEEE | 2022


    Traffic police gesture recognition and judgment based on Openpose

    Zhu, Qingxiang / Dang, Xueming | SPIE | 2024


    Traffic Police Gesture Recognition by Pose Graph Convolutional Networks

    Fang, Zhijie / Zhang, Wuqiang / Guo, Zijie et al. | IEEE | 2020


    TRAFFIC POLICE GESTURE RECOGNITION BY POSE GRAPH CONVOLUTIONAL NETWORKS

    Fang, Zhijie / Zhang, Wuqiang / Guo, Zijie et al. | British Library Conference Proceedings | 2020