Performance Improvement of Human Voice Recognition System using Gaussian Mixture Model

Abstract: The voice recognition system is most prominent technique of identification of human voice. This is used for security purposes. Recent research concentrates on developing systems that would be much more robust against variability in environment, speaker and language. Hence today’s researches mainly focus on ASR systems with a large vocabulary that support speaker independent operation with continuous speech in different languages. The performance of Speaker recognition systems has improved due to recent advances in speech processing techniques but there is still need of improvement. In this paper we present the recognition performance of numeric digits and speech alphabets. The Gaussian mixture model is used for classification of the speech signal. To improve, we use hybrid concentration of Mel Frequency Cepstral Coefficients (MFCC) and Linear predictive coding (LPC). The entire coding was done in MATLAB and the system was tested for its reliability.

Keywords: Feature extraction, feature matching, MFCC, LPC, Gaussian mixture model (GMM)