Generalized gamma density-based score functions for fast and flexible ICA

In this contribution, we propose an entirely novel family of flexible score functions for blind source separation (BSS), based on the family of generalized gamma densities. To blindly extract the independent source signals, we resort to the popular FastICA approach, whilst to adaptively estimate the parameters of such score functions, we use an efficient method based on maximum likelihood (ML). Experimental results with sources employing a wide range of statistical distributions, indicate that the proposed flexible FastICA (FF-ICA) technique significantly outperforms conventional independent component analysis (ICA) methods, which operate only on a fixed score function regime.