Overlapped sub-band modulation spectrum normalization techniques for robust speech recognition.
Hao-Teng FanWei-Jeih YehJeih-weih HungPublished in: FSKD (2013)
Keyphrases
- speech recognition
- subband
- noisy environments
- speech processing
- multiresolution
- hidden markov models
- language model
- speech synthesis
- discrete wavelet transform
- wavelet transform
- high frequency
- image compression
- speech recognition technology
- speech recognizer
- wavelet coefficients
- filter bank
- automatic speech recognition
- speaker identification
- speech recognition systems
- pattern recognition
- speech signal
- low frequency
- low pass
- neural network
- speaker independent
- wavelet decomposition
- bit rate
- feature vectors
- image processing
- multiscale
- speaker verification
- image data
- signal processing