Image Regularization with Morphological Gradient Priors Using Optimal Structuring Elements for Each Pixel

Abstract

As an image prior for image restoration, the use of the sum of the morphological gradient for an image has been proposed. In this paper, we propose a method of optimizing the structuring element (SE) for each pixel, in particular, by greatly reducing the computational burden for optimization by adopting a new evaluation method for the SE in simulated annealing. By optimizing the SE for each pixel, edges of the image can be faithfully evaluated, and the improvement of restoration accuracy can be expected. An experimental result shows the effectiveness of the proposed method.