Cluster Representation of the Structural Description of Images for Effective Classification

The problem of image recognition in the computer vision systems is being studied. results development efficient classification methods, given figure processing speed, based on analysis segment representation structural description form a set descriptors are provided. We propose three versions classifier according to following principles: “object–etalon”, “object descriptor–etalon” and “vector object–etalon”, which not similar level integration researched data analysis. options for constructing clusters over whole descriptions etalon database, separately each etalons, as well optimal method compare sets centers etalons object, implemented. An experimental rating efficiency created classifiers terms productivity, time, quality has been realized applied. proposed methods classify without error. have formed inference about approaches centers. time developed hundreds times less than traditional one, reducing accuracy.