Voice recognition through the use of Gabor transform and heuristic algorithm

Abstract

Increasingly popular use of verification methods based on specific characteristics of people like eyeball, fingerprint or voice makes inventing more accurate and irrefutable methods of that urgent. In this work we present the voice verification based on Gabor transformation. The proposed approach involves creation of spectrogram, which serves as a habitat for the population of selected heuristic algorithm. The use of heuristic allows for the features extraction to enable identity verification using classical neural network. The results of the research are presented and discussed to show efficiency of the proposed methodology.