A Spectral Clustering Approach to Underdetermined Post-Nonlinear Blind Source Separation of Sparse Sources

Steven Van Vaerenbergh, Student Member, IEEE, Ignacio Santamaría, Senior Member, IEEE
Dept. of Communications Engineering, University of Cantabria, Spain
E-mail: {steven,nacho}@gtas.dicom.unican.es
Telephone: 34-942201392-13, 34-942201552
Fax: 34-942201488

Abstract:

This letter proposes a clustering-based approach for solving the underdetermined (i.e. fewer mixtures than sources) post-nonlinear blind source separation (PNL BSS) problem when the sources are sparse. Although various algorithms exist for the underdetermined BSS problem for sparse sources, as well as for the PNL BSS problem with as many mixtures as sources, the nonlinear problem in an underdetermined scenario has not been satisfactorily solved yet. The method proposed in this work aims at inverting the different nonlinearities, thus reducing the problem to linear underdetermined BSS. To this end, first a spectral clustering technique is applied that clusters the mixture samples into different sets corresponding to the different sources. Then, the inverse nonlinearities are estimated using a set of multilayer perceptrons (MLPs) that are trained by minimizing a specifically designed cost function. Finally, transforming each mixture by its corresponding inverse nonlinearity results in a linear underdetermined BSS problem, which can be solved using any of the existing methods.

Index Terms

Blind source separation, underdetermined source separation, post-nonlinear mixtures, sparse sources, spectral clustering, multilayer perceptrons.