Rebrun I.A., Borodina N.I. Automatic evaluation of the indistinct attri bu tes of graphic objects.

Abstract: In this article the authors discuss the ways to automate the growing method of diagnosis – a tezigraphy. It has long been known that biological ß uids crystallize when dried. Under pathological conditions, the crystallization properties of biological ß uids change. As a result of the studies it has been found that this method is very sensitive and can be used on a preclinical stage to diagnose the viral and infectious diseases. To determine the nature of the disease it is necessary to analyze the obtained images of biocrystals to describe their shape and determine the way of distribution by the number of attributes. The article gives the algorithms and methods for automated evaluation of the image attri butes on the image samples obtained during the tezigraphy researches. The article inves tigates the segmentation algorithm based on edge detection using Þ ltering algorithm for calculating the areas and detecting attributes of form and algorithms of modeling and fractal analysis in the consideration of tree and other complex crystallogramms. The authors pre sents a program, which allows to obtain esti mates of the distribution in area, perimeter and other characteristics of the crystals, as well as statistical data for a variety of crystals grown on the same surface.

Keywords: