Blind separation of mixed images using higher order statistics

The separation of mixed images is a very exciting area of research, especially when no a priori information is available about the mixed images. In such a case, blind signal separation (BSS) technique is needed to recover independent sources given only sensor observations, which are assumed to be unknown independent images. Classical eigen and singular value decomposition techniques, which are based on second order statistics, fail to blindly separate mixed signals in many circumstances. On the other hand, many techniques based on higher order statistics were suggested to blindly separate mixed independent signals and images. These techniques are generally known as independent component analysis (ICA) as they are oriented towards searching for the statistically independent components in the signal mixture (instead pf the uncorrelated components in the techniques based on second order statistics). In this paper, we investigate the performance of a technique based on higher order statistics, namely higher order eigen-value decomposition (HOEVD) in the blind separation of mixed images. Illustrative examples using different types of images and different mixing matrices are utilized to demonstrate the ability of the considered technique.