Sectioning in image processing

For reasons of mathematical tractability, simplifying assumptions are usually made about the statistical nature of images that are to be processed. Among these assumptions are stationarity of the signal and noise processes and the space invariance of the blur. These assumptions rarely hold over a large image, but can be considered accurate over small subsections of an image. With this idea in mind, it is natural to divide an image into small, possibly overlapping, sections and process each section with consideration to its peculiar characteristics. This technique was used with success on problems involving iterative numerical image restorations, signal-dependent noise, and space-variant point spread functions