Handbook of Blind Source Separation

Chapter 6

Iterative algorithms

V. Zarzoso and A. Hyvärinen

6.1 Introduction

As seen in the preceding chapters of this book, the problems of blind source separation (BSS) and independent component analysis (ICA) can be solved by optimizing a so-called contrast function measuring at the separator output a known property of the sources or independent components. Specific approaches differ in how this optimization is carried out. The present chapter focuses on techniques performing the contrast optimization in an iterative fashion. An iterative technique is defined by an equation specifying how the separating filters are updated at each step of the algorithm. The update rule modifies the extracting filters so as to approach an optimum value of the contrast ...