Maximum likelihood sub-band weighting for robust speech recognition.
Donglai ZhuSatoshi NakamuraKuldip K. PaliwalRen-Hua WangPublished in: INTERSPEECH (2003)
Keyphrases
- speech recognition
- subband
- maximum likelihood
- noisy environments
- speaker adaptation
- speech synthesis
- hidden markov models
- speech signal
- speaker identification
- multiresolution
- speech processing
- filter bank
- language model
- speech recognizer
- wavelet transform
- pattern recognition
- frequency domain
- low frequency
- automatic speech recognition
- speech recognition technology
- speech recognition systems
- image compression
- discrete wavelet transform
- wavelet coefficients
- high frequency
- isolated word
- wavelet decomposition
- expectation maximization
- low pass
- bit rate
- speaker independent
- feature vectors
- additive noise
- mixture model
- em algorithm
- feature selection
- machine learning