Dual-Compression Neural Network with Optimized Output Weighting For Improved Single-Channel Speech Enhancement.
Stefan ThaleiserAleksej ChinaevRainer MartinGerald EnznerPublished in: IWAENC (2022)
Keyphrases
- single channel
- speech enhancement
- neural network
- multi channel
- noisy environments
- independent component analysis
- noise reduction
- frequency domain
- signal to noise ratio
- wavelet decomposition
- prior information
- image compression
- wiener filter
- signal processing
- brain activity
- pattern recognition
- linear prediction
- high frequency
- background noise
- multiresolution