Space Invariant Deconvolution Using Maximum A Posteriori (MAP) Estimation for Imaging Inverse Problem

Abstract

Coupling edge and texture detection using directional priors along with blind image deconvolution using an iterative alternating minimization algorithm gives an efficient deblurring. By assuming correct directional priors for edge detection in a constrained Maximum A Posteriori Estimation (MAP) accurate PSF estimation is achieved. The results are compared with existing blind deconvolution algorithms and are found to be more efficient.