Speaker adaptation based on combination of MAP estimation and weighted neighbor regression.
Lei HeJian WuDitang FangWenhu WuPublished in: ICASSP (2000)
Keyphrases
- map estimation
- maximum a posteriori
- maximum likelihood
- speaker adaptation
- markov random field
- noisy images
- maximum a posterior
- expectation maximization
- em algorithm
- machine learning
- energy function
- denoising
- higher order
- model selection
- generative model
- speech recognition
- bayesian networks
- markov networks
- image segmentation
- image processing
- feature selection