Класифікація зображень на підставі кластерування

Abstract

The result of an experimental study of image classification methods, which have a high calculation speed using data centers in the form of cluster centroids, is presented. As an alternative to the traditional method based on linear search, classification methods with clustering of descriptions within the entire base of reference images and separately for each reference image have been introduced. The simulation results showed the expediency of using the clustering apparatus to speed up classification.