Regularization methods for blind deconvolution and blind source separation problems (Q5959638)

Blind deconvolution is the identification of a point spread function and an input signal from an observation of their convolution, while blind source separation is the recovery of a vector of input signals from a vector of observed signals which are mixed by an unknown linear operator. This paper is devoted to these two problems. In particular, the authors show that both problems are paradigms of nonlinear ill-posed problems, and consequently, regularization techniques have to be used for stable numerical reconstructions. The authors develop a rigorous convergence analysis for regularization techniques for the solution of the blind deconvolution and the blind separation problems. Convergence of regularized point spread functions and signals to a solution is established, and a convergence rate result depending on the noise level is presented. Moreover, the authors prove convergence of the alternating minimization algorithm for the numerical solution of regularized blind deconvolution problems and present some numerical examples. Finally, they show that many neural network approaches for blind inversion can be considered in the framework of regularization theory.