Blind separation of any source distributions via high-order statistics

Authors:
Manal Taoufiki;Abdellah Adib;Driss Aboutajdine
Affiliations:
LRIT-Faculté des Science de Rabat, Av. Ibn Battouta, BP 1014, Rabat, Morocco;DPG-Institut Scientifique, Av. Ibn Battouta, BP 703, Rabat, Morocco and LRIT-Faculté des Science de Rabat, Av. Ibn Battouta, BP 1014, Rabat, Morocco;LRIT-Faculté des Science de Rabat, Av. Ibn Battouta, BP 1014, Rabat, Morocco
Venue:
Signal Processing
Year:
2007

Citing 6
Cited 4

Independent component analysis, a new concept?

Signal Processing - Special issue on higher order statistics
Reference-based blind source separation using a deflation approach

Signal Processing
Quadratic Higher Order Criteria for Iterative Blind Separation of a MIMO Convolutive Mixture of Sources

IEEE Transactions on Signal Processing
A blind source separation technique using second-order statistics

IEEE Transactions on Signal Processing
Independent component analysis and (simultaneous) third-ordertensor diagonalization

IEEE Transactions on Signal Processing
A generalization of joint-diagonalization criteria for sourceseparation

IEEE Transactions on Signal Processing

A new behavior of higher order blind source separation methods for convolutive mixture

Digital Signal Processing
Fast frequency template matching using higher order statistics

IEEE Transactions on Image Processing
"Weighting for more": Enhancing characteristic-function based ICA with asymptotically optimal weighting

Signal Processing
Denoising by second order statistics

Signal Processing

Quantified Score

Hi-index	0.08

Visualization

Abstract

An effective technique to perform instantaneous blind source separation by considering a reformulation of the referenced contrast (RC) [A. Adib, E. Moreau, D. Aboutajdine, Blind sources separation by simultaneous generalized referenced contrasts diagonalization, in: ICA, Nara, Japan, 2003, pp. 657-661] is presented. This novel method applies to separate all sources having any distributions including Gaussian ones. Our proposition allows one to assess the capabilities of the RC to break the obviousness that high-order statistics-based methods are restricted to the mixtures allowing at most one Gaussian source. Our design constrains the reference signals choice to avoid the cancellation of the cross-cumulants between this latter and the separator outputs. Simulation studies are presented to support the potential of the approach in terms of source separation.