Abstract
In this paper, the restoration problem of blurred images which are blurred by the point spread function with Gauss distribution. Usually, it is difficult to presume the original image if the blurred image does not have something a special condition. Therefore, original images in this paper is limited to binary images. To restore the original image, we introduce a new function which evaluates a binary degree of images. This evaluation function is defined by using Otsu's method which is the technique of making to binarization of the gray scale image. Restoration of the original image is enabled by presuming standard deviation of Gauss distribution by using this function. By computer simulation, it was shown that the original image took the maximum of the evaluation function. Hence, if this evaluation function is used, it is confirmed to be able to restore the original image.