Speech enhancement by LSTM-based noise suppression followed by CNN-based speech restoration.
Maximilian StrakeBruno DefraeneKristoff FluytWouter TirryTim FingscheidtPublished in: EURASIP J. Adv. Signal Process. (2020)
Keyphrases
- speech enhancement
- noise suppression
- noise reduction
- noise removal
- median filter
- wiener filter
- impulsive noise
- denoising algorithm
- signal to noise ratio
- gaussian noise
- noisy environments
- impulse noise
- denoising
- image restoration
- speech signal
- edge preserving
- noise level
- image deblurring
- filtering algorithm
- edge detection
- extensive simulations
- single channel
- vocal tract
- image denoising
- linear prediction
- speech recognition
- image processing
- image enhancement
- noisy images
- additive noise
- wavelet domain
- anisotropic diffusion
- multi channel
- color images