Total generalized variation and shearlet transform based Poissonian image deconvolution

摘要

Integrating the advantages of total generalized variation and shearlet transform, this article introduces a hybrid regularizers scheme for deconvolving Poissonian image. Computationally, a highly efficient alternating minimization algorithm associated with variable splitting approach is described to obtain the optimal solution in detail. Illustrationally, in comparison with several current state-of-the-art numerical methods, numerical simulations consistently demonstrate the outstanding performance of our proposed approach to deblurring Poissonian image, in terms of both restoration accuracy and feature-preserving ability.