Dynamic Minimum Subband Spectral Subtraction and Its Application in Robust Speech Recognition.
Xin MaYuhua PengPublished in: ICICIC (3) (2006)
Keyphrases
- speech recognition
- subband
- spectral subtraction
- speech signal
- filter bank
- noisy environments
- frequency band
- image compression
- wavelet transform
- multiresolution
- hidden markov models
- bit rate
- wavelet decomposition
- linear prediction
- speech synthesis
- language model
- pattern recognition
- high frequency
- wavelet coefficients
- automatic speech recognition
- discrete wavelet transform
- noisy speech
- feature vectors
- wavelet domain
- frequency domain
- information retrieval
- wavelet packet
- low frequency
- speaker identification
- speaker verification
- multiscale
- speech enhancement
- image segmentation