Probabilistic models of digital region maps based on Markov random fields with short- and long-range interaction

Abstract We introduce novel generating probabilistic models representing a raster region map, i.e. an image with nominal scale of signal values (region labels), as sample realization of a Markov random field of labels with Gibbs joint probability distribution. The models take into account various kinds of local interaction of labels in adjacent and displaced pixels forming the nearest and more distant neighborhoods of each pixel. Examples of the maps obtained by pixelwise stochastic relaxation show high expressive abilities of these models to describe textural images.