Markov model based noise modeling and its application to noisy speech recognition using dynamical features of speech.
Tetsunori KobayashiRyuji MineKatsuhiko ShiraiPublished in: ICASSP (2) (1994)
Keyphrases
- noisy speech
- speech enhancement
- speech recognition
- speech signal
- noisy environments
- signal to noise ratio
- hidden markov models
- spectral subtraction
- background noise
- feature space
- feature extraction
- noise reduction
- automatic speech recognition systems
- vocal tract
- structural features
- low level
- feature vectors
- additive noise
- linear prediction
- automatic speech recognition
- single channel
- spectral features
- speech synthesis
- wavelet coefficients
- non stationary