Speaker Recognition through Effectively Configured Artificial Neural Network
Abstract
The purpose of work to recognize the speaker's voice through Artificial Neural Network (ANN) in associate with Particle Swarm Optimization (PSO) for identifying appropriate weights for ANN. The preliminary research extracts feature from the signal with the aid of Linear Prediction-filter Coefficients (LPC) and Mel-Frequency Cepstral Coefficient (MFCC). The extracted features can be used as input parameters for the ANN technique; subsequent, the research aims to identify appropriate weights for performance enhancement. Determining the appropriate weights through manual or trial-and-error process consumes a long time, to resolve the study incorporates optimization techniques. The performance of employed techniques evaluates through various measures, the outcome evident the superiority of the anticipated technique over contest techniques.