Text independent speaker recognition system using GMM

Abstract

The idea of the Speaker Recognition Project is to implement a recognizer which can identify a person by processing his/her voice. The basic goal of the project is to recognize and classify the speeches of different persons. This classification is mainly based on extracting several key features like Mel Frequency Cepstral Coefficients (MFCC's) from the speech signals of those persons by using the process of feature extraction method. The above features may consist of pitch, amplitude, frequency etc. Using a statistical model like Gaussian mixture model (GMM) and features extracted from those speech signals we build a unique identity for each person who enrolled for speaker recognition. Estimation and Maximization algorithm are used, an elegant and powerful method for finding the maximum likelihood solution for a model with latent variables, to test the later speakers against the database of all speakers who enrolled in the database. Use of Fractional Fourier Transform for feature extraction is also suggested improving the speaker recognition efficiency.