Image Regularization Using Total Variation and Morphological Gradient Priors with Optimization of Structuring Element
Shoya Oohara, Mitsuji Muneyasu, Soh Yoshida, Makoto Nakashizuka · 2018
As an image prior for image restoration, the sum of morphological gradients for an image has previously been proposed. Optimization of the structuring element (SE) used for this morphological gradient using a genetic algorithm (GA) has also been proposed. This method uses the minimized value of the objective function of the restoration problem as the fitness of the GA. However, this value does not necessarily coincide with an objective evaluation such as the mean square error. Therefore, in this paper, we formulate the objective function using the morphological gradient and total variation as a new image prior for an image restoration problem. The proposed objective function makes it possible to almost match the fitness to the objective evaluation and can improve the restoration accuracy. It also solves the problem of the artifact due to the unsuitability of the SE for the image. An experiment shows the effectiveness of the proposed image restoration method.