An ICA based approach for blind deconvolution of three-dimensional signals

The paper presents a new algorithm based on independent component analysis (infomax approach) for blind deconvolution of three-dimensional signals. In this sense, it represents an extension of what has been done, so far, for one and two dimensional signals. Such a method has also been successfully implemented in the frequency domain to alleviate the computational burden, due to the efficient calculation of convolutive terms. Experimental tests carried out on synthetic source signals, with sub-Gaussian and super-Gaussian distributions, confirm the validity of the idea.