Time-frequency structured decorrelation of speech signals via nonseparable Gabor frames.
Werner KozekHans Georg FeichtingerPublished in: ICASSP (1997)
Keyphrases
- speech signal
- linear prediction
- blind source separation
- wavelet transform
- speech recognition
- automatic speech recognition
- vocal tract
- short time fourier transform
- background noise
- automatic speech recognition systems
- non stationary
- hidden markov models
- feature extraction
- face recognition
- signal processing
- speech enhancement
- gabor filters
- frame rate
- sound signals
- filter bank
- noisy environments
- pattern recognition
- independent component analysis
- feature space
- spectral analysis
- speech quality
- multiscale
- multi frame
- key frames
- video frames
- neural network
- feature vectors
- video sequences
- high resolution
- image coding