Discriminative spectral-temporal multiresolution features for speech recognition.
Philip McMahonNaomi HarteSaeed VaseghiPaul M. McCourtPublished in: ICASSP (1999)
Keyphrases
- speech recognition
- multiresolution
- speech recognition systems
- cepstral coefficients
- hidden markov models
- feature extraction
- automatic speech recognition
- pattern recognition
- speech synthesis
- language model
- speaker independent
- speech signal
- speech recognizer
- speech processing
- speaker adaptation
- speech recognizers
- speaker dependent
- low level
- classification accuracy
- wavelet transform
- subband
- discriminative training
- spectral features
- noisy environments
- signal to noise ratio
- feature vectors
- speech retrieval
- feature set