ANALYSIS OF EMG SIGNALS USING WAVELET ANALYSIS AND SOME FEATURES

Abstract

ElectroMyoGram (EMG) is a method used to measure the problems in muscles and nerve cells of the body. Comparison of overall EMG waveform pattern and shape enables doctors to diagnose possible diseases. Currently there is computer based analysis which employs certain signal processing to diagnose a patient based on EMG recording. Signal processing usually takes the form of a transformation of a signal into another signal that is in some sense more desirable than the original. The purpose of this research is to help in identifying the Normal, Myopathy and Neuropathy signal using the method of Discrete Wavelet Transform (DWT) and various classifiers which includes k-Nearest neighbour (kNN) approach, Artificial Neural Network (ANN) Classifier and Support Vector Machine (SVM) Classifier. DWT coefficients are used to extract the relevant information from the EMG input data which are Energy, Mean and Standard Deviation values. Then the extracted features data is analyzed and classified using the classifiers such as k-Nearest neighbour (kNN) approach, Artificial Neural Network (ANN) Classifier and Support Vector Machine (SVM) Classifier. The proposed algorithm is implemented and also tested in MATLAB software. The EMG signal are being selected and tested from PhysioNet Database using MIT-BIH Database. The Classifier (SVM) used successfully classifies the Normal, Myopathy and Neuropathy signals with the rate of accuracy as 95.55%. The analysis system also can be achieved using rest of the classifiers such as kNN and ANN with accuracies of 73.33%,88.88% respectively for each sample tested of Normal, Myopathy and Neuropathy classes proposed.