Kernel PCA Based Estimation of the Mixing Matrix in Linear Instantaneous Mixtures of Sparse Sources.
Frédéric DesobryCédric FévottePublished in: ICASSP (5) (2006)
Keyphrases
- mixing matrix
- blind source separation
- source separation
- source signals
- blind separation
- independent component analysis
- nonlinear functions
- principal component analysis
- non stationary
- simple linear
- independent components
- denoising
- support vector
- single channel
- negative matrix factorization
- sparse representation
- factor analysis
- kernel function
- feature space
- discriminant analysis
- dimensionality reduction
- random projections
- input output
- blind deconvolution
- signal processing
- neural network