SIGN LANGUAGE RECOGNITION BASED ON ELECTROMYOGRAPHY (EMG) SIGNAL USING ARTIFICIAL NEURAL NETWORK (ANN)

Abstract

EMG signal is a muscular signal. It can represent equally time & frequency domain. EMG provides measurement of muscular performance. Hence it is a direct representation of strength of health. It is used for identification & treatment tasks by comparing with normal person. EMG biofeedback can be instantaneous as compare to manual inspection. EMG signal used in management of prostheses & to improve the effectiveness of rehabilitation robotics. The motive of this paper is to describe the process of detecting different predefined hand gestures using artificial neural network (ANN).This different hand gestures used by deaf & dumb people to understand the things. The EMG sample signatures are extracted from the signals for all movement and then given to ANN for training & classification. A back-propagation (BP) network has been used for the recognition of sign, as it works well for different biosignal. The features like autoregression coefficient, fast Fourier transform, short time Fourier transform, wavelet transform coefficient, mean absolute value, root mean square, variance, standard deviation, Mean frequency, zero crossing and slope sign change are selected to train the neural network.