Electromyography Feature Extraction, Selection and Classification: A REVIEW

Abstract

Inpast decades remarkable progress has done in processing, extraction of feature and classification
of motion of EMG signals.EMG Signals are termed as recording in terms of electric measurement
that can be measured from the surface of skin which carry important and useful information
regarding activity of muscle. How much muscle have been contracted i.e. in terms of strength can be
monitored and captured by activity of muscle through skin surface extraction. EMG signal analysis is
one of the main procedures to identify action, posture and gestures. These signals are basically
complex by nature, with the utilization of proper signal processing tools, it can be classified muscular
activities which are differentiated. In this paper we provide a comprehensive review of EMG Signal
its feature selection & extraction and classification methods. Wavelet transforms and various
classifiers such as PCA, LDA and SVM has been reviewed to classify and analyze the EMG Signal in
an appropriate way. We generally first review the overall basis of EMG signals and its various
processing stages. Finally the paper has been concluded with small discussion on recent challenges
and future work.