On the solution of circular and noncircular complex KLD-ICA in the presence of noise

This paper aims at studying the solution of linear independent component analysis (ICA) based on Kullback-Leibler divergence (KLD) for a linear noisy mixing model in the determined case. The derivation is done using a perturbation analysis valid for small noise variance. We study the noncircular complex and the circular complex case. We show that for a wide range of both the shape parameter and the noncircularity index of the generalized Gaussian distribution (GGD), the signal-to-interference-plus-noise ratio (SINR) of KLD-based ICA is close to that of linear minimum mean squared error (MMSE) estimation.