Kernel Machines Beat Deep Neural Networks on Mask-based Single-channel Speech Enhancement.
Like HuiSiyuan MaMikhail BelkinPublished in: CoRR (2018)
Keyphrases
- speech enhancement
- single channel
- kernel machines
- neural network
- multi channel
- semi supervised
- support vector
- independent component analysis
- noisy environments
- noise reduction
- wiener filter
- wavelet decomposition
- frequency domain
- pattern recognition
- prior information
- signal to noise ratio
- brain activity
- speech signal
- linear prediction
- signal processing
- wavelet transform
- speech recognition
- support vector machine
- image restoration