sEMG-based gesture recognition becomes popular in human-computer interaction due to its non-invasive nature and ease of signal acquisition. However, electrode-shift due to long-time use or re-wearing severely decreases the gesture recognition accuracy. In this paper, we apply a LightViT network to sEMG-based gesture recognition to improve the recognition accuracy under electrode-shift conditions with economical computational complexity. Firstly, an sEMG signals database is established to simulate conditions that the electrode-band in forearm shift between 0°–45°. Secondly, a short-time Fourier transform is used to transform sEMG signals into two-dimensional images, preserving the temporal dependency and correlation of the original time series. Then, the LightViT network is applied to feature extraction and classification of two-dimensional sEMG signals. A concept of feature alignment in transfer learning is introduced by applying Triplet Loss and Focal Loss as loss functions in the model to pay more attention to complex samples. Experiment results shows that the LightViT network achieves a gesture recognition accuracy of 91.9% within a range of 0°–45° electrode-shift compared with ResNet (85.7%) and mViT (89.3%) and demonstrates fewer methods parameters and floating-point operations.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Gesture Recognition Based on LightViT Network Against sEMG Electrode Shift


    Additional title:

    Lect.Notes Computer


    Contributors:
    Yang, Huayong (editor) / Liu, Honghai (editor) / Zou, Jun (editor) / Yin, Zhouping (editor) / Liu, Lianqing (editor) / Yang, Geng (editor) / Ouyang, Xiaoping (editor) / Wang, Zhiyong (editor) / Liu, Tao (author) / Zhang, Zheng (author)

    Conference:

    International Conference on Intelligent Robotics and Applications ; 2023 ; Hangzhou, China July 05, 2023 - July 07, 2023



    Publication date :

    2023-10-21


    Size :

    10 pages




    Type of media :

    Article/Chapter (Book)


    Type of material :

    Electronic Resource


    Language :

    English




    Gesture Recognition Based on LightViT Network Against sEMG Electrode Shift

    Liu, Tao / Zhang, Zheng / Li, Kairu | TIBKAT | 2023


    Real-time Gesture Recognition Based on Improved Artificial Neural Network and sEMG Signals

    Zhang, Wenzhe / Shuai, Liguo / Kan, Haoxuan | British Library Conference Proceedings | 2021


    SEMG Based Recognition for Lumbar Muscle Fatigue During Prolonged Driving

    Tao, Xin / Cheng, Bo / Wang, Bo et al. | Springer Verlag | 2012


    SEMG Based Recognition for Lumbar Muscle Fatigue During Prolonged Driving

    Tao, Xin / Cheng, Bo / Wang, Bo et al. | Tema Archive | 2012


    D-S evidential theory on sEMG signal recognition

    Ding, Weiliang / Li, Gongfa / Sun, Ying et al. | British Library Online Contents | 2017