A Statistically Principled and Computationally Efficient Approach to Speech Enhancement Using Variational Autoencoders.
Manuel ParienteAntoine DeleforgeEmmanuel VincentPublished in: INTERSPEECH (2019)
Keyphrases
- computationally efficient
- speech enhancement
- noisy environments
- noise reduction
- denoising
- signal to noise ratio
- wiener filter
- single channel
- speech signal
- linear prediction
- vocal tract
- background noise
- image segmentation
- smoothing algorithm
- edge detection
- filter bank
- multi frame
- multi channel
- additive noise
- speech recognition
- optical flow
- automatic speech recognition
- computational complexity
- neural network
- non stationary
- frequency domain
- computer vision