1. An enhanced speaker recognition system using a combined approach of speech signal and egg signal
Authors : Kunal Anand, Subhasish Mohapatra
Pages : 211-221
Keywords : Speaker Recognition (SR), MFCC, EGG, GMM, Excitation source Abstract :
Speaker Recognition (SR) is the process of detecting a speaker based on the features obtained from the speaker’s voice sample. The traditional SR system is based on the recorded speech signal as the only source of information. The decision to authenticate or reject the speaker is taken using the MFCC features extracted from the speech sample obtained from the speaker. This is done using Gaussian Mixture Model (GMM). However, this speech signal approach has few limitations like it performs well in a noiseless environment, but the performance degrades in noisy environments due to the effect of noise on the speech signal. An alternate approach to overcome the limitation of speech signal approach is the SR system using excitation source characteristics using Electroglottography (EGG) signal. This paper is intended to do a comparative study of the performance of SR system using speech signal and EGG system, along with discussing the benefits and shortcomings of each approach to suggest a
Citing this Journal Article :Kunal Anand, Subhasish Mohapatra, "An enhanced speaker recognition system using a combined approach of speech signal and egg signal ", Volume 6 Issue 3 - January 2016, 211-221
Click here to Submit Copyright Takedown Notice for this article.