Enhancement and recognition of noisy speech within an autoregressive hidden Markov model framework using noise estimates from the noisy signal.
Beth T. LoganAnthony J. RobinsonPublished in: ICASSP (1997)
Keyphrases
- noisy speech
- hidden markov models
- speech recognition
- noisy environments
- speech enhancement
- background noise
- speech signal
- autoregressive
- automatic speech recognition systems
- signal to noise ratio
- non stationary
- linear prediction
- noise reduction
- spectral subtraction
- conditional random fields
- pattern recognition
- speaker identification
- probabilistic model
- single channel
- automatic speech recognition
- low contrast
- random fields
- speech synthesis
- machine learning
- language model
- generative model
- image sequences
- image processing