Histogram-based subband powerwarping and spectral averaging for robust speech recognition under matched and multistyle training.
Mark HarvillaRichard M. SternPublished in: ICASSP (2012)
Keyphrases
- speech recognition
- subband
- noisy environments
- filter bank
- wavelet transform
- bit rate
- image compression
- wavelet coefficients
- speech signal
- linear prediction
- multiresolution
- isolated word
- hidden markov models
- wavelet decomposition
- high frequency
- language model
- feature vectors
- speech recognizer
- cepstral coefficients
- automatic speech recognition
- wavelet domain
- frequency domain
- pattern recognition
- discrete wavelet transform
- low frequency
- wavelet packet
- spectral analysis
- acoustic models
- image coding
- image coder
- speaker adaptation
- speaker identification
- speech recognition systems
- image denoising
- neural network
- speaker independent
- video coding
- video signals
- computational complexity
- computer vision