Iterations of a non-linear transformation for enhancement of digital images

Abstract A novel non-linear transformation for sharpening digitized gray-level images is defined. The transformation replaces the gray level value at a point by either the minimum or the maximum of gray levels in its neighborhood, the choice depending on which one is closer in value to the original gray level. It is shown that after a finite number of iterations, the image stabilizes to one in which every point is either a local maximum or else a local minimum. The ability of several iterations of the transformation to restore white-black crispness in a fuzzy image is shown both analytically and by example.