基于深度学习与灰度图的肌电机械手手势识别  

Electromyography(EMG)Robotic Hand Gesture Recognition Based on Deep Learning and Grayscale Images

作　　者：罗肖媛 黄文静 曾学武 段小刚 郑哲文 王子怡 LUO Xiao-yuan;HUANG Wen-jing;ZENG Xue-wu;DUAN Xiao-gang;ZHENG Zhe-wen;WANG Zi-yi(School of Materials Science and Engineering,Central South University of Forestry&Technology,Changsha Hunan 410004;Hetian(Hunan)International Engineering Management Co.,Ltd,Changsha Hunan 410006;Central South Intelligence Hunan Industry 4.0 Innovation Center,Changsha Hunan 410000)

机构地区：[1]中南林业科技大学材料科学与工程学院,湖南长沙410004 [2]和天(湖南)国际工程管理有限公司,湖南长沙410006 [3]中南智能湖南工业4.0创新中心,湖南长沙410000

出　　处：《长沙航空职业技术学院学报》2023年第4期33-38,共6页Journal of Changsha Aeronautical Vocational and Technical College

基　　金：湖南省自然科学基金面上项目“基于概率模糊理论的肌电假肢手抓握模式与力解码研究”(编号:2022JJ31015)阶段性研究成果。

摘　　要：为了提高肌电机械手手势识别的精度,首先将肌电信号信息转换为图片,实现更直观、更明显的肌电信息展示,然后使用迁移学习微调Res Net50模型,构建基于肌电信号灰度图的肌电机械手手势识别模型,并使用该模型对三种手势进行识别实验。结果显示,ResNet50模型对基于肌电灰度图的手势识别率高达93%,有效减少了过拟合现象并提高了识别精度。This study aims to convert electromyography(EMG)signals into two-dimensional grayscale images and use grayscale images and deep learning to improve the accuracy of EMG-controlled robotic hand gesture recognition.Firstly,the EMG signal information is transformed into images to achieve a more intuitive and clear display of EMG information.Then,transfer learning is used to fine-tune the ResNet50 model,constructing an EMG-controlled robotic hand gesture recognition model based on EMG grayscale images.This model is used to recognize three types of hand gestures,and the experimental results show that the ResNet50 model achieves a recognition rate of 93%for EMG grayscale-based gesture recognition.This effectively reduces overfitting and improves recognition accuracy.

关 键 词：肌电信号 灰度图像 卷积神经网络 手势肌电识别 

分 类 号：S776.222[农业科学—森林工程] TP24[农业科学—林学]