Blind Deconvolution and Source Separation: Recent Results and Open Issues
Phillip A Regalia · 2002
Abstract The problems of blind de convolution and blind signal separation have met with fruitful interplay in recent years. The two problems share a common setting: One is given a mixture of independent random variables or ‘sources,’ where the mixture may he temporal (as assumed in de convolution) or spatial (as assumed in signal separation), and the goal is to restore one of the sources. A multi-source convolutional mixture setting combines elements of both problems, and can be specialized to either one depending on the spatial and temporal characteristics of the mixture.