Algoritmus pro automatickou obrazovou detekci sítnicových lézí

Abstract

The aim of this bachelor thesis is to create an automatic algorithm for visual detection of retinal lesions and consequently scrutinising their geometric properties and relations between size and size of optical disc. This algorithm is being designed for easier classification of rethinopathy of prematurity. The main idea is to detect retinal lesions and their detachment from the background. The algorithm follows basic methods of image processing in the MATLAB enviroment. Constituent part of this thesis is also to design a user friendly graphical interface for easy use of the algorithm. Testing of the algorithm has been approved on real medical data. The data were provided by Fakultní nemocnice Ostrava.