Flexible Independent Component Analysis

This paper addresses an independent component analysis (ICA) learning algorithm with exible nonlinearity, so named as exible ICA, that is able to separate instantaneous mixtures of suband super-Gaussian source signals. In the framework of natural Riemannian gradient, we employ the parameterized generalized Gaussian density model for hypothesized source distributions. The nonlinear function in the exible ICA algorithm is controlled by the Gaussian exponent according to the estimated kurtosis of demixing lter output. Computer simulation results and performance comparison with existing methods are presented. .