Modeling attention zones on the basis of an analysis of local features of the image texture

摘要

In the basic model developed and published earlier by the authors, the information-content function proposed by Muchnik and Zavalishin is transformed in such a way that the attention zone (AZ) is fixed at the centers of gravity (CGs) of the objects of an image distinguished from the surroundings by the local brightness. This makes it possible to directly recognize images subjected to an affine transformation characteristic of the shapes of distant surfaces projected on the retina. Unlike the basic model described in that article, the expanded model analyzes more general attributes of an image-the local features of its texture. This allows our approach to be used to analyze a broader class of images. The final goal of forming AZs is to automatically decompose and structurally analyze the images of actual three-dimensional scenes. (C) 2008 Optical Society of America.