Speech Dereverberation Based on Probabilistic Models of Source and Room Acoustics.
Tomohiro NakataniBiing-Hwang JuangKeisuke KinoshitaMasato MiyoshiPublished in: ICASSP (1) (2006)
Keyphrases
- probabilistic model
- speech recognition
- graphical models
- generative model
- hidden variables
- language model
- bayesian inference
- expectation maximization
- latent variables
- speech signal
- bayesian networks
- mixture model
- conditional random fields
- database
- spoken dialogue systems
- multi stream
- endpoint detection
- speaker identification
- emotion recognition
- multiple sources
- control system
- learning algorithm