Speech enhancement based on stationary bionic wavelet transform and maximum a posterior estimator of magnitude-squared spectrum.
Talbi MouradPublished in: Int. J. Speech Technol. (2017)
Keyphrases
- maximum a posterior
- wavelet transform
- speech enhancement
- map estimation
- least squares
- maximum a posteriori
- image coding
- linear prediction
- noise reduction
- speech signal
- non stationary
- single channel
- markov random field
- noisy environments
- wavelet decomposition
- maximum likelihood
- multiresolution
- signal to noise ratio
- multiscale
- subband
- image compression
- wavelet coefficients
- wavelet domain
- high frequency
- vocal tract
- wiener filter
- background noise
- smoothing algorithm
- image restoration
- feature extraction
- energy function
- multi channel
- additive noise
- higher order
- filter bank
- edge detection
- signal processing
- speech recognition
- pairwise
- pattern recognition
- similarity measure
- image sequences