DefinePK

Abstract

Identification of person voice from their characteristics is said to be speaker recognition; it is generally utilize in telecommunication, voice biometrics, criminal investigations and various industrial applications. To pursue, Mel-Frequency Cepstral Coefficient (MFCC) and Linear Prediction-filter Coefficients (LPC) are utilizes to extract features from voice signal as preliminary process. Subsequently, these features employ to Artificial Neural Network (ANN) to recognize speaker. This research includes configuring conventional ANN structure holding single hidden layer associate with ten neurons. To conserve time and process complexity optimization techniques incorporates to identify appropriate configuration for performance enhancement. An optimization technique includes Evolutionary Algorithm (EA), Genetic Algorithm (GA), Particle Swarm Optimization (PSO) and Monarch Butterfly Optimization (MBO). The comparative techniques are K Nearest Neighbour (KNN), Random Forest (RF), Radial Basis Neural Network (RBNN), Back-Propagation Neural Network (BPNN) and Generalized Regression Neural Network (GRNN). The result reveals the performance of MBO in configuring ANN accomplishes 99% accuracy in real time database and 99.8% in benchmark database which is superior results over contest techniques.