The Effect of GMM Order and CMS on Speaker Recognition with Reverberant Speech

Author

Shabtai, Noam R. ; Zigel, Yaniv ; Rafaely, Boaz

Author_Institution

Dept. of Electr. & Comput. Eng., Ben-Gurion Univ. of the Negev, Beer-Sheva

fYear

2008

Firstpage

144

Lastpage

147

Abstract

Speaker recognition is used today in a wide range of applications. The presence of reverberation, in hands-free systems for example, results in performance degradation. The effect of reverberation on the feature vectors and its relation to optimal GMM order are investigated. Optimal model order is calculated in terms of minimum BIC and KIC, and tested for EER of a GMM-based speaker recognition system. Experimental results show that for high reverberation time, reducing model order reduces EER values of speaker recognition. The effect of CMS on state of the art GMM and AGMM-based speaker recognition systems is investigated for reverberant speech. Results show that high reverberation time reduces the effectiveness of CMS.

Keywords

Bayes methods; Gaussian processes; adaptive signal processing; cepstral analysis; error statistics; reverberation; speaker recognition; Bayesian information criterion; GMM order; Kullback information criterion; adaptive Gaussian mixture model; cepstral mean subtraction; equal error rate; hands-free system; reverberant speech; speaker recognition; Biomedical engineering; Cepstral analysis; Collision mitigation; Degradation; Feature extraction; Microphone arrays; Reverberation; Speaker recognition; Speech; System testing; Gaussian mixture model; Speaker recognition; cepstral mean subtraction; model order; reverberation;

fLanguage

English

Publisher

ieee

Conference_Titel

Hands-Free Speech Communication and Microphone Arrays, 2008. HSCMA 2008

Conference_Location

Trento

Print_ISBN

978-1-4244-2337-8

Electronic_ISBN

978-1-4244-2338-5

Type

conf

DOI

10.1109/HSCMA.2008.4538707

Filename

4538707