Towards end-to-end speech enhancement with a variational U-Net architecture.
Eike Jannik NustedeJörn AnemüllerPublished in: CoRR (2020)
Keyphrases
- end to end
- speech enhancement
- admission control
- signal to noise ratio
- noisy environments
- noise reduction
- speech signal
- image segmentation
- congestion control
- single channel
- application layer
- pattern recognition
- multiresolution
- image restoration
- speech recognition
- high quality
- real time
- computer vision
- wiener filter
- neural network