Noisy BSS based on Joint Diagonalization of Differences of Correlation Matrices

A. Tanaka, H. Imai, and M. Miyakoshi (Japan)


blind source separation, second-order-statistics, joint diag onalization, stationary noise.


The aim of blind source separation is to recover mutually independent unknown source signals from observations ob tained through an unknown linear mixture system. In many existing methods, it is assumed that the observations are not contaminated by an observation noise. Although methods for a noisy observation model are proposed, assumptions for the noise, such as Gaussianity, spatially or temporally whiteness, may limit the application area of these methods. In this study, we propose a novel blind source separation method for a noisy observation model, in which only sta tionarity is imposed on the noise, based on a joint diago nalization of differences of correlation matrices.

Important Links:

Go Back