Journals Proceedings

Abstract

The domain area of this topic is Bio-metric. Speaker Recognition is biometric system. This paper deals with speaker recognition by HMM (Hidden Markov Model) method. The recorded speech signal contains background noise. This noise badly affects the accuracy of speaker recognition. Discrete Wavelet Transforms (DWT) greatly reduces the noise present in input speech signal. DWT often outperforms as compared to Fourier Transform, due to its capability to represent the signal precisely, in both frequency & time domain. Wavelet thresholding is applied to separate the speech and noise, enhancing the speech consequently. The system is able to recognize the speaker by translating the speech waveform into a set of feature vectors using Mel Frequency Cepstral Coefficients (MFCC) technique. Hidden Markov Model (HMM) provides a highly reliable way for recognizing a speaker. Hidden Markov Models have been widely used, which are usually considered as a set of states with Markovian properties and observations generated independently by those states. With the help of Viterbi decoding most likely state sequence is obtained. This state sequence is used for speaker recognition. For a database of size 50 in noisy environment, obtained result is 96% which is better than previous methods used for speaker recognition.