Sciweavers

Share
ICASSP
2011
IEEE

Rapid speaker adaptation with speaker adaptive training and non-negative matrix factorization

10 years 10 months ago
Rapid speaker adaptation with speaker adaptive training and non-negative matrix factorization
In this paper, we describe a novel speaker adaptation algorithm based on Gaussian mixture weight adaptation. A small number of latent speaker vectors are estimated with non-negative matrix factorization (NMF). These base vectors encode the correlations between Gaussian activations as learned from the train data. Expressing the speaker dependent Gaussian mixture weights as a linear combination of a small number of base vectors, reduces the number of parameters that must be estimated from the enrollment data. In order to learn meaningful correlations between Gaussian activations from the train data, the NMF-based weight adaptation was combined with vocal tract length normalization (VTLN) and feature-space maximum likelihood linear regression (fMLLR) based speaker adaptive training based. Evaluation on the 5k closed and 20k open vocabulary Wall Street Journal tasks shows a 4% relative word error rate reduction over the speaker independent recognition system which already incorporates VTL...
Xueru Zhang, Kris Demuynck, Hugo Van hamme
Added 20 Aug 2011
Updated 20 Aug 2011
Type Journal
Year 2011
Where ICASSP
Authors Xueru Zhang, Kris Demuynck, Hugo Van hamme
Comments (0)
books