ABSTRACT
Of the two primary approaches to the classic source separation problem, only one does not impose potentially unreasonable model and likelihood constraints: the Bayesian statistical approach. Bayesian methods incorporate the available information regarding the model parameters and not only allow estimation of the sources and mixing coefficients, but
TABLE OF CONTENTS
part I|1 pages
Fundamentals
part II|1 pages
Models
part III|1 pages
Generalizations