Eigenspace-based maximum a posteriori linear regression for rapid speaker adaptation.
Kuan-Ting ChenHsin-Min WangPublished in: ICASSP (2001)
Keyphrases
- linear regression
- speaker adaptation
- maximum a posteriori
- maximum likelihood
- map estimation
- least squares
- em algorithm
- hyperparameters
- speech recognition
- posterior distribution
- expectation maximization
- mixture model
- bayesian framework
- markov random field
- kernel regression
- input image
- regression methods
- automatic speech recognition
- image reconstruction
- prior distribution
- principal component analysis
- semi supervised
- pattern recognition