One approach to decoding the multizonal images

In this paper, a new approach to decoding multizonal images is described. In the first case, the decoding of each component of a colored image is performed irrespective of the rest of the components. In the second case, all the components of a colored image are replaced by one monochromatic image. It should be noted that in both cases the complete or partial loss of a very important decoding attribute occurs which is the dependence between random components describing the pixel properties. To eliminate this, the concept of a random distance is used as the measure of likeness of real-scene elements. It is introduced and used for searching the zones of interest and segmentation.