Cluster analysis based on quality of representative estimations of digital images

In this article we introduce images quality indicators, the methods of their evaluation, and the most representative indicators are selected Then we carry out the cluster analysis of sample images for subsequent use of he results for the training signal obtaining required for fuzzy controller knowledge base design. The fuzzy conroller is nedeed to manage parameters of the pattern recognition algorithm. Cluster analysis is performed using the k-means algorithm and Gaussian mixture model.