A Comparison of Session Variability Compensation Techniques for SVM-Based Speaker Recognition

McLaren, Mitchell L. and Vogt, Robert J. and Baker, Brendan J. and Sridharan, Sridha (2007) A Comparison of Session Variability Compensation Techniques for SVM-Based Speaker Recognition. In Proceedings Interspeech 2007, pages pp. 790-793, Antwerp, Belgium.

Full text available as: PDF (Author version) - Requires Adobe Acrobat Reader or other PDF viewer.

Abstract

This paper compares two of the leading techniques for session variability compensation in the context of GMM mean supervector SVM classifiers for speaker recognition: inter-session variability modelling and nuisance attribute projection. The former is incorporated in the GMM model training while the latter is employed as a modified SVM kernel. Results on both the NIST 2005 and 2006 corpora demonstrate the effectiveness of both techniques for reducing the effects of session variation. Further, system- and score-level fusion experiments show that the combination of the two methods provides improved performance.

Item Type:	Conference Paper
Status:	Published
Subjects:	280000 Information, Computing and Communication Sciences > 280200 Artificial Intelligence and Signal and Image Processing
ID Code:	7594
Deposited By:	Vogt, Robert
Deposited On:	25 February 2008
Alternative Locations:	http://www.interspeech2007.org/, http://www.isca-speech.org
Copyright Owner:	Copyright 2007 (please consult author)
Additional Information:	For more information, please refer to the conference’s website (see hypertext link) or contact the author.