The problem of pattern classification at the action of hindrances is discussed. The model of processing of characteristic features based on the histogram analysis of the data base is suggested. The conversion to the space of histogram features simplifies the computation of the descriptions conformity and shortens the time of classification. Results of computer experiments that confirms efficiency of suggested approach are shown.
images classification, structural methods, characteristic features, descriptor, histogram features, measure of similarity
"Modely obrabotky deskryptorov kharakternыkh pryznakov yzobrazhenyi na osnove analyza hystohramm" ,
Information Processing Systems,