Identifying vessels in the eye using image processing methods

This thesis is concerned with the extraction of the vessels using image processing methods for early diagnosis and treatment of diseases and to be able to easily monitor changes in the blood vessels in the eye in the retina due to various diseases. The aim of the thesis is a complete extraction of the blood vessels for the early diagnosis and treatment of diseases of the retina. The method of identifying shape line has been used in the application carried out for the thesis in which blood vessels in retinal images are determined. Firstly, in the method of determining the shape line, pixel intensity values that have the same color and contours are collected in a vector matrix. Then, the area of the contours located within this vector matrix are calculated, and those areas too small to be vessels are eliminated. Vessels are identified by drawing the remaining contours after areas that are not suitable for being vessel are eliminated. Various methods are used to extract vessels. Many programs used for vessel extraction methods are usually written in Matlab code. Speed is important in applications. For that reason, the upper level programming languages are preferred while developing an application [1]. In this study, an open source image processing library OpenCV has been written by applying the method of vessel extraction. The filters in OpenCV image processing library that are ready to be used have been applied to eye images taken with fundus camera. After that, the results have been analyzed.