A method and system for surface electromyography signal classification based on convolutional neural network

Abstract

The invention discloses a surface electromyographic signal classification method and system based on a convolutional neural network. The method includes: preprocessing the electromyographic signals of each channel; extracting feature information from the preprocessed electromyographic signals of each channel; extracting a variety of feature information from each channel; The feature information is combined to generate multiple sets of two-dimensional data; each set of two-dimensional data contains only one type of feature information; a convolutional neural network model is constructed; the convolutional neural network model includes convolutional layers, pooling layers, ReLu layers, and full connections layer and softmax layer; train the convolutional neural network model through multiple sets of two-dimensional data; classify the surface EMG signal through the trained convolutional neural network model, and identify gestures. The present invention improves the classification and recognition accuracy of surface electromyography signals by extracting feature information from multi-channel electromyography signals, expanding data dimensions, and taking advantage of the high-precision classification advantage of the convolutional neural network.

Description

technical field [0001] The invention relates to the field of electromyographic signal classification, in particular to a surface electromyographic signal classification method and system based on a convolutional neural network. Background technique [0002] Surface EMG is a bioelectrical signal acquired on the skin surface during human movement, and it has broad application prospects in human-machine interface, rehabilitation medicine, and game entertainment. Among them, action classification and recognition based on EMG signals is an important link in these applications, and it is also a prerequisite for EMG signals to be widely used. [0003] At present, the action classification methods based on EMG signals mainly include the steps of signal acquisition, signal preprocessing, feature information extraction and classification. Among them, feature information extraction is usually carried out by manual feature extraction method based on experience; the classification step ...

AI Technical Summary

Problems solved by technology

However, due to the low spatial resolution of the action classification based on sparse EMG signals, the data cannot be well converted into the data format of the convolutional neural network, and it still faces the problem of low classification and recognition accuracy.

Images