Speaker and Session Variability in GMM-Based Speaker Verification

We present a corpus-based approach to speaker verification in which maximum-likelihood II criteria are used to train a large-scale generative model of speaker and session variability which we call joint factor analysis. Enrolling a target speaker consists in calculating the posterior distribution of...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on audio, speech, and language processing speech, and language processing, 2007-05, Vol.15 (4), p.1448-1460
Main Authors: Kenny, P., Boulianne, G., Ouellet, P., Dumouchel, P.
Format: Article
Language:English
Subjects:
Adaptation model
Applied sciences
Availability
Error analysis
Exact sciences and technology
Factor analysis
Gaussian mixture
Information, signal and communications theory
Large-scale systems
Mathematical models
Natural language processing
NIST
Performance analysis
Performance evaluation
Signal processing
Speaker recognition
speaker verification
Speech processing
Speech recognition
Statistical analysis
Statistical distributions
Statistics
Studies
Telecommunications and information theory
Testing
Trains
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We present a corpus-based approach to speaker verification in which maximum-likelihood II criteria are used to train a large-scale generative model of speaker and session variability which we call joint factor analysis. Enrolling a target speaker consists in calculating the posterior distribution of the hidden variables in the factor analysis model and verification tests are conducted using a new type of likelihood II ratio statistic. Using the NIST 1999 and 2000 speaker recognition evaluation data sets, we show that the effectiveness of this approach depends on the availability of a training corpus which is well matched with the evaluation set used for testing. Experiments on the NIST 1999 evaluation set using a mismatched corpus to train factor analysis models did not result in any improvement over standard methods, but we found that, even with this type of mismatch, feature warping performs extremely well in conjunction with the factor analysis model, and this enabled us to obtain very good results (equal error rates of about 6.2%)
ISSN:1558-7916
2329-9290
1558-7924
2329-9304
DOI:10.1109/TASL.2007.894527