Feature Extraction in Electromyography by Digital Signal Processing Techniques

Abstract

Myoelectric signals are the electrical manifestation associated with the movements exerted by the muscular system in the mamal beings (including the human). Examination of these signals should reveal the Status of the muscles as well as the driving nervous system. This is important in diagnosis as well as prosthesis for the health of mankind and aids for handicaped. This would not be possible unless powerful digital processing techniques are available. In this paper, several techniques are investigated so as to extract the features of the ME signals both in time and frequency domains. The extracted features are subsequently employed in an automatic diagnostic classification system to decide whether or not they correspond to a normal muscle.