Abstract

Info

Title:: Recognition of hand gestures with gender differences based on surface electromyographic signals

Author(s):: LEI Jianchao; LIU Dongbo; FANG Yu; ZHUANG Zujiang; LIU Junhao; School of Electrical Engineering and Electronic Information, Xihua University, Chengdu 611730, China

Keywords:: Keywords: hand gesture; surface electromyographic signal; energy compensation; wavelet packet decomposition

Abstract:: Abstract: For the same gestures, there are some gender differences in surface electromyographic signals (sEMG). Herein a method combining sliding average energy and energy compensation is proposed to reduce the gender differences for the recognition of hand gestures. The sEMG of 10 hand gestures are collected. The active segment is detected by sliding average energy, and then the energy of female motion segment is compensated. Five wavelet functions of wavelet packet decomposition, namely Db4, Bior3.2, Haar, Sys8 and Dmey, were used to extract features. Finally, the obtained data are classified and recognized by particle swarm optimization-support vector machine. The results show that energy compensation improves the identification of features, reduces gender differences, and increases the recognition rate of hand gestures.