Enhancement of speech degraded by coherent and incoherent noise using a cross-spectral estimator.
Régine Le Bouquin-JeannèsAhmad Akbari AziraniGérard FauconPublished in: IEEE Trans. Speech Audio Process. (1997)
Keyphrases
- noisy environments
- spectral features
- speech signal
- linear predictive coding
- speech enhancement
- least squares
- estimation error
- image enhancement
- figure of merit
- noisy data
- random noise
- linear prediction
- additive noise
- speech recognition
- missing data
- maximum likelihood
- noise model
- linear minimum mean square error
- audio visual
- degraded images
- image processing
- speech synthesis
- single channel
- background noise
- video sequences
- spectral analysis
- median filter
- noise reduction
- signal to noise ratio
- impulsive noise
- spoken language
- image noise
- broadcast news
- dialogue system
- automatic speech recognition
- image contrast
- multi channel
- bias variance
- image quality