Back propagation feed forward neural network approach for Speech Recognition
Neelima Rajput, SANI VERMA · 2014
In this paper a biologically motivated approach for the English alphabet speech recognition is implemented by using a self-organized neural network. The designing of an accurate and effective speech recognition system is a challenging task in the area of human computer interface. Linear Predictive coding (LPC) is used for learn Feature extraction of input audio signals. Back propagation (BP) is a feed forward neural network and it propagates the error in backward direction to update the weights of hidden layers. The error is difference of actual output and target output computed on the basis of gradient descent method. The performance of the system is evaluated on the basis of recognition rate. We have used BP neural network architecture to recognize the time varying input data. The proposed provides better accurate results than the existing systems for the English Alphabet speech recognition.