A novel speech enhancement approach based on Singular Value Decomposition and Genetic Algorithm.
Amin ZehtabianHamid HassanpourShahrokh ZehtabianVicente ZarzosoPublished in: SoCPaR (2010)
Keyphrases
- singular value decomposition
- speech enhancement
- noisy environments
- noise reduction
- signal to noise ratio
- single channel
- singular values
- least squares
- low rank
- vocal tract
- speech signal
- dimensionality reduction
- principal component analysis
- multi channel
- smoothing algorithm
- singular vectors
- wiener filter
- background noise
- motion estimation
- neural network
- linear prediction
- non stationary
- signal processing
- feature extraction