Speaker Identification with Counter Propagation Neural Network

Speaker recognition has been an active area of research in the past due to its diverse applications and it continues to be a challenging research topic. Counter Propagation Neural networks provides an attractive possibilities for solving signal processing & pattern classification problems. Several algorithms have been proposed for choosing the network prototypes & training the network. The proposed thesis implements a novel speaker recognition system using counter propagation neural network and LPC coefficients. The proposed work is tested on the speaker Identification problem. Features are obtained by using Liner Predictive coding (LPC) Coefficients and these features are classified by using counter propagation neural network. The efficiency of the proposed method is tested on the standard TIMIT dataset. It is shown that the use of LPC coefficients results in better performance in terms of percentage of correct classification.