Work is devoted to the problem of images large databases analysis in terms of solving the problem of intelligent search. To solve this problem it is convenient to introduce the set of all images in the form of some classes as their similarities. For this unit may apply cluster analysis methods and which are discussed in this paper. In this matrix proposed two modifications of known approaches to simplify the image analysis by eliminating operations vectoring-devektoring of source data.
images, matrix hierarchical clustering, sample, cluster