Alternating split Bregman method for the bilaterally constrained image deblurring problem

摘要

This paper studies the image deblurring problem based on a bilateral constraint by convexly combining two classes of total-variation-type functionals. The proposed model including two L-1-norm terms leads to some numerical difficulties, so we employ the alternating split Bregman method (ASB) to solve it which can be reinterpreted as Douglas-Rachford splitting applied to the dual problem. We also prove that the alternating split Bregman method owns the convergence rate O(1/M) for the iteration M. Experimental results demonstrate the viability and efficiency of the proposed model and algorithm to restore blurring and noisy images. (C) 2014 Elsevier Inc. All rights reserved.