paper: "Multiresolution framework for blind source separation" (fwd)

[Michael Zibulevsky]

Announcing a paper ...

Title:   Multiresolution framework for blind source separation

Authors: P. Kisilev, M. Zibulevsky, Y.Y. Zeevi, B.A. Pearlmutter 

ABSTRACT:

The concern of the blind source separation problem is to extract the
underlying source signals from a set of their linear mixtures, where the
mixing matrix is unknown. It was discovered recently, that use of sparsity
of sources in some signal dictionary dramatically improves the quality of
separation. In this work we use the property of multiscale transforms,
such as wavelet or wavelet packets, to decompose signals into sets of
local features with various degrees of sparsity. We use this intrinsic
property for selecting the best (most sparse) subsets of features for
further separation. Experiments with simulated signals, musical sounds
and images demonstrate further significant improvement of separation
quality.

URL of the ps file:

http://ie.technion.ac.il/~mcib/multisepMP9a.ps.gz

Contact: paulk at tx.technion.ac.il  mzib at ee.technion.ac.il