WaveCRN: An Efficient Convolutional Recurrent Neural Network for End-to-end Speech Enhancement.
Tsun-An HsiehHsin-Min WangXugang LuYu TsaoPublished in: CoRR (2020)
Keyphrases
- end to end
- recurrent neural networks
- speech enhancement
- neural network
- feed forward
- noisy environments
- hidden layer
- recurrent networks
- speech signal
- noise reduction
- signal to noise ratio
- congestion control
- back propagation
- vocal tract
- network architecture
- linear combination
- background noise
- single channel
- image compression
- edge detection