Parameter estimation in total variation blind deconvolution

In this paper we present a methodology for parameter estimation in total variation (TV) blind deconvolution. By formulating the problem in a Bayesian framework, the unknown image, blur and the model parameters are simultaneously estimated. The resulting algorithms provide approximations to the posterior distributions of the unknowns by utilizing variational distribution approximations. We show that some of the current approaches towards TV-based blind deconvolution are special cases of our formulation. Experimental results are provided to demonstrate the performance of the algorithms.