Frame-Weighted Bayes Factor Scoring for Speaker Verification
作者： Robert Vogt , Subramanian Sridharan
DOI:
关键词: Machine learning 、 Frame (networking) 、 Context (language use) 、 NIST 、 Speaker recognition 、 Weighting 、 Computer science 、 Artificial intelligence 、 Bayesian probability 、 Bayes factor 、 Adaptation (computer science)
摘要: In this paper, the Bayes factor is considered as a replacement verification criterion to likelihood-ratio test in context of GMM-based speaker verification. An advantage Bayesian method that it allows for incorporation prior information and uncertainty parameter estimates into scoring process, complementing adaptation used training. A development factors GMMs presented based on incremental well-suited inclusion existing GMM-UBM systems. This extended include weighting frames account their statistical dependencies. Experiments 1999 NIST Speaker Recognition Evaluation corpus demonstrate improved performance over expected log-likelihood ratio scoring. These findings are supported with results from modified version Extended Data 2003.
参考文章(11)
Chin-Hui Lee, Jean-Luc Gauvain, Bayesian Adaptive Learning and Map Estimation of HMM Springer, Boston, MA. pp. 83- 107 ,(1996) , 10.1007/978-1-4613-1367-0_4
Subramanian Sridharan, Robert Vogt, Bayes factor scoring of GMMs for speaker verification. Odyssey 04. pp. 173- 178 ,(2004)
Douglas A. Reynolds, Comparison of background normalization methods for text-independent speaker verification. conference of the international speech communication association. ,(1997)
Ganesh N. Ramaswamy, Jir ´ i Navratil, THE AWE AND MYSTERY OF T-NORM conference of the international speech communication association. ,(2003)
Alvin Martin, Mark Przybocki, The NIST 1999 Speaker Recognition Evaluation An Overview Digital Signal Processing. ,vol. 10, pp. 1- 18 ,(2000) , 10.1006/DSPR.1999.0355
Roland Auckenthaler, Michael Carey, Harvey Lloyd-Thomas, Score Normalization for Text-Independent Speaker Verification Systems Digital Signal Processing. ,vol. 10, pp. 42- 54 ,(2000) , 10.1006/DSPR.1999.0360
Hui Jiang, K. Hirose, Qiang Huo, Robust speech recognition based on a Bayesian prediction approach IEEE Transactions on Speech and Audio Processing. ,vol. 7, pp. 426- 440 ,(1999) , 10.1109/89.771309
Hui Jiang, Li Deng, A Bayesian approach to the verification problem: applications to speaker verification IEEE Transactions on Speech and Audio Processing. ,vol. 9, pp. 874- 884 ,(2001) , 10.1109/89.966090
Sridha Sridharan, Jason W. Pelecanos, Feature Warping for Robust Speaker Verification Proceedings of 2001 A Speaker Odyssey: The Speaker Recognition Workshop. pp. 213- 218 ,(2001)
David G. Stork, Richard O. Duda, Peter E. Hart, Pattern Classification ,(1973)