Unsupervised speaker adaptation based on HMM sufficient statistics in various noisy environments.
Shingo YamadeAkinobu LeeHiroshi SaruwatariKiyohiro ShikanoPublished in: INTERSPEECH (2003)
Keyphrases
- noisy environments
- speaker adaptation
- speech recognition
- sufficient statistics
- automatic speech recognition
- speaker independent
- hidden markov models
- speaker identification
- speech signal
- digit recognition
- speech recognizer
- speaker verification
- speech enhancement
- pattern recognition
- speech synthesis
- data points
- language model
- unsupervised learning
- k means
- maximum likelihood
- supervised learning
- expectation maximization
- broadcast news
- semi supervised
- hidden variables
- noise reduction
- neural network
- machine learning
- denoising
- speech recognition systems
- multiscale