Adaptive training with noisy constrained maximum likelihood linear regression for noise robust speech recognition.
D. K. KimM. J. F. GalesPublished in: INTERSPEECH (2009)
Keyphrases
- noisy environments
- speech recognition
- linear regression
- maximum likelihood
- speech signal
- least squares
- wall street journal corpus
- automatic speech recognition
- speaker adaptation
- language model
- isolated word
- speaker identification
- background noise
- regression methods
- linear predictors
- pattern recognition
- hidden markov models
- noise reduction
- acoustic models
- speech enhancement
- additive noise
- speech processing
- speech synthesis
- discriminative training
- speech recognizer
- nonlinear regression
- noisy speech
- training set
- speech recognition technology
- speech recognition systems
- speaker independent
- word recognition
- speech recognizers
- expectation maximization
- neural network
- gaussian distribution
- probabilistic model