ScaleFormer: Transformer-based speech enhancement in the multi-scale time domain.
Tianci WuShulin HeHui ZhangXueliang ZhangPublished in: APSIPA ASC (2023)
Keyphrases
- speech enhancement
- multiscale
- noise reduction
- single channel
- noisy environments
- signal to noise ratio
- frequency domain
- edge detection
- speech signal
- linear prediction
- multi channel
- vocal tract
- background noise
- scale space
- smoothing algorithm
- wavelet decomposition
- image segmentation
- natural images
- wiener filter
- wavelet transform
- image processing
- edge preserving
- spatial domain
- image coding
- wavelet domain
- image denoising
- optic flow
- lossless compression
- wavelet coefficients
- speech recognition
- subband
- image compression