Melnyk G. Information Technology of Structural Textures Analysis and Synthesis in Histological Image Processing Automated Systems

The object of study: the processes of analysis and synthesis of histological images. Objective: To develop information technology for analysis and synthesis of structural textures based on the theory of crystallographic groups to increase automation level of histological images analysis. Research methods: theory of algorithms, theory of crystallographic groups, theory of pattern recognition and analytical geometry, structural and object-oriented programming technologies. Scientific novelty of the work. For the first time the method of evaluation of histological images structural atypia is developed. It is based on the theory of crystallographic groups and affine transformations of a tree and allowed to determine the plane symmetry group for the individual structures of microobjects and quantify the degree of distortion. For the first time the image synthesis method is developed. It is based on parametric texture synthesis and automatic control of the color features, and allowed to expand the training sample during classification procedure development. The method of texture segmentation of histological images was further developed. It is based on histogram analysis of the geometrical feature-moments, thus improving the quality of segmentation. Applied information technology was created on the basis of the developed methods and tools. It enables to solve the problem of the numerical evaluation of structural tissue atypia of histological images and create histological images test sample.