Fourth-order criteria for blind sources separation
Ali Mansour, Christian Jutten · IEEE Transactions on Signal Processing · 1995
Various criteria based on contrast functions as well as higher order statistics are used for solving the problem of blind separation of sources. In the paper, for the case of instantaneous mixtures of two sources, it is proved that the minimization or the cancellation of a really simple criterion, a fourth-order cross-cumulant, leads to a set of solutions whose spurious ones can be simply cancelled by using a decorrelation.>