Speech enhancement using blind source separation and two-channel energy based speaker detection.
Erik M. VisserTe-Won LeePublished in: ICASSP (1) (2003)
Keyphrases
- blind source separation
- speech signal
- speech enhancement
- speech recognition
- single channel
- independent component analysis
- vocal tract
- automatic speech recognition
- noisy environments
- multi channel
- non stationary
- noise reduction
- linear prediction
- image processing
- hidden markov models
- language model
- neural network
- signal processing
- signal to noise ratio
- image denoising
- multiresolution
- wavelet transform
- pattern recognition
- edge detection