Filter bank subtraction for robust speech recognition.
Kazuo OnoeHiroyuki SegiTakeshi KobayakawaShoei SatoToru ImaiAkio AndoPublished in: INTERSPEECH (2002)
Keyphrases
- speech recognition
- filter bank
- noisy environments
- image coding
- computationally efficient
- subband
- hidden markov models
- language model
- multiresolution
- multiscale
- pattern recognition
- speech signal
- perfect reconstruction
- frequency domain
- automatic speech recognition
- speech synthesis
- speaker identification
- speech recognition technology
- speech recognizer
- spectral analysis
- speech recognition systems
- wavelet packet
- signal processing
- neural network
- speaker independent
- wavelet filters
- computational complexity
- wavelet transform
- maximum likelihood