Low-Latency Single Channel Speech Dereverberation Using U-Net Convolutional Neural Networks.
Ahmet E. BulutKazuhito KoishidaPublished in: INTERSPEECH (2020)
Keyphrases
- low latency
- single channel
- convolutional neural networks
- speech enhancement
- multi channel
- high throughput
- high speed
- independent component analysis
- highly efficient
- virtual machine
- stream processing
- real time
- speech recognition
- speech signal
- wiener filter
- brain activity
- frequency domain
- noise reduction
- prior information
- wavelet decomposition
- noisy environments
- signal to noise ratio
- signal processing
- wavelet transform
- vocal tract
- feature vectors
- moving objects