Diabetic retinopathy is one of the most frequent complications of diabetes, which leads to severe consequences, including rapid and irreversible vision loss. The laser coagulation procedure to treat diabetic retinopathy consists in applying a series of microburns to the fundus to deal with macular edema. The existing hardware/software packages are primarily based on the use of a predetermined pattern for coagulate arrangement. However, due to the complex form of edema and vascular system, this approach leads to an uneven arrangement. To solve the problem, we propose a new approach based on the application of sphere packing algorithms (circle packing in two-dimensional images) in the specified area of interest. Since one of the main requirements for the laser coagulation procedure is that it should have the minimum duration, a problem of the computational complexity of the developed algorithms arises. This requirement is completely ignored by the existing approaches, therefore the development of new high-performance coagulate arrangement algorithms is highly relevant. In the paper, we propose seven new algorithms for coagulate arrangement and provide a detailed analysis of key characteristics of the algorithms. The characteristics considered have made it possible to extract information relating to the algorithm properties. Regularity is determined by the median, asymmetry and the kurtosis; determinism is determined by the variance and the mean.

No abstract

The computer diagnostic system of eye diseases is considered. To improve the quality of diagnostics we propose an algorithm for the informative features formation, using methods of discriminant analysis. The method for receiving an informtiveness estimation is described. The research confirming the efficiency of the formed features for classification of images of an fundus was conducted by means of classification by support vector machine. The algorithm possesses a sufficient level of universality and may be applied to increase the informtiveness of any feature set.