Comon P., Jutten C. (eds.) Handbook of Blind Source Separation. Independent Component Analysis and Applications

Academic Press, 2010. — 840 p.In signal processing, a generic problem consists in separating a useful signal from noise and interferences. Classical approaches of the twentieth century are based on a priori hypotheses, leading to parameterized probabilistic models. Blind Source Separation (BSS) attempts to reduce these assumptions to the weakest possible.
As shown in this handbook, there are various approaches to the BSS problem, depending on the weak a priori hypotheses one assumes. The latter include either statistical independence of source signals or their scarcity, among others.
This handbook is an extension of another book which appeared in 2007 in French, and published by Hermes. The present version contains more chapters and many additions, provided by contributors with international recognition. It is organized into 19 chapters, covering all the current theoretical approaches, especially Independent Component Analysis, and applications. Although these chapters can be read almost independently, they share the same notations and the same subject index. Moreover, numerous cross-references link the chapters to each other.Information.
Contrasts.
Algebraic methods after prewhitening.
Iterative algorithms.
Second-order methods based on color.
Convolutive mixtures.
Algebraic identification of under-determined mixtures.
Sparse component analysis.
Quadratic time-frequency domain methods.
Bayesian approaches.
Non-negative mixtures.
Nonlinear mixtures.
Semi-blind methods for communications.
Overview of source separation applications.
Application to telecommunications.
Biomedical applications.
Audio applications.