A two-dimensional image model based on occlusion and maximum entropy

This paper provides new insights into the formation of two-dimensional image autocorrelation functions. We model an image as a maximum-entropy composition of individual occluding object images that have random positions, shapes and intensities. We derive the autocorrelation function of this image model, give an example, and comment on the reasonableness of the frequently-made assumptions of autocovariance separability and isotropy.