Abstract
Application of stochastic geometry methods to pattern recognition is analyzed. The discourse is based on trace transformations of original images introduced by (Fedotov, 1990) into images on the Mobius band. Based on the new geometric transformation, a new approach towards the construction of features, independent of image motions or their linear transformations, is put forward. A prominent characteristics of the group of features under consideration is representing each of them as a consecutive composition of three functionals. Such a structure of features helps automatically generate a greater number of new constructive signs of recognition. Such a powerful shift of stress from the decision procedures onto features' mass application, accounts for speaking about a new interpretation of images, and enhancing intellectual power of recognition systems.

This publication has 1 reference indexed in Scilit: