ICCRN: Inplace Cepstral Convolutional Recurrent Neural Network for Monaural Speech Enhancement.
Jinjiang LiuXueliang ZhangPublished in: ICASSP (2023)
Keyphrases
- recurrent neural networks
- speech enhancement
- linear prediction
- noisy environments
- speech recognition
- speech signal
- image coding
- feed forward
- prediction error
- neural network
- noise reduction
- complex valued
- signal to noise ratio
- lossless compression
- recurrent networks
- echo state networks
- reservoir computing
- hidden layer
- network architecture
- artificial neural networks
- single channel
- autoregressive
- subband
- vocal tract
- spectral analysis
- multiscale
- automatic speech recognition
- multi channel
- background noise
- wiener filter
- back propagation
- linear combination
- machine learning