A spectrally-based signal subspace approach for speech enhancement.
Yariv EphraimHarry L. Van TreesPublished in: ICASSP (1995)
Keyphrases
- speech enhancement
- signal subspace
- noisy environments
- noise reduction
- signal to noise ratio
- linear array
- correlation matrix
- single channel
- sensor array
- speech signal
- hyperspectral
- linear prediction
- estimation error
- vocal tract
- eigenvalue decomposition
- synthetic aperture radar
- covariance matrix
- wiener filter
- multi channel
- additive noise
- non stationary
- edge preserving
- multiscale
- sar images
- speech recognition
- multiresolution