A Model of Recognition Algorithms Based on Threshold Functions and Object Proximity Assessment
摘要
The authors consider the construction of a model of recognition algorithms for solving the problems of classifying the objects represented in a high-dimensional feature space. A new approach to developing such a model is proposed based on constructing a set of representative features and determining the corresponding set of three-dimensional threshold functions in the process of generating an extreme recognition algorithm. A structural description of the proposed model of recognition algorithms in the form of a sequence of computational procedures is given. The parameterization of these algorithms has been carried out, which makes it possible to set and solve the problem of determining an extreme recognition algorithm within the limits of the created model. The results of a comparative analysis of the proposed and known recognition algorithms are given.