The Mathematical Model of a Visual System for Object Recognition Against Equally Bright Backgrounds

Abstract

Currently, there is a growing interest in the problem of object recognition in various practical applications. Despite the obvious ease with which a human recognizes the objects surrounding him or her, no visual system models have hitherto been developed that would solve the recognition problem in the same general scope and with the same accuracy. There are various approaches to solving the object recognition problem. One of possible approaches is to build a mathematical model that would adequately describe the human response to the objects observed. The complexity of applying a mathematical model to the human eye lies in the fact that the algorithm according to which a human recognizes objects is still unknown. In the study presented, the main hypothesis is the assumption that under threshold recognition conditions, an observer, as in the case of detecting objects, makes the maximum use of a priori and a posteriori information, trying to solve the problem assigned to him or her in the best way. This allows an assumption to be made that the observer characteristics should be close to those of an optimal image receiver. The article discusses the operation of an optimal image receiver in regard of modeling the characteristics of the human visual system in solving the problem of recognizing objects against equally bright backgrounds. Two cases are considered: recognition of pairs of objects and recognition of any object from a set of other objects known to the observer. The main calculation relations describing the response of observers to the images presented are obtained, and the further line of work is indicated.