The role of automated systems is growing nowadays, the same tendency is showing up in the medical sector: different diagnostic software systems are applied increasingly. There are several kind of images taken from patients such as x-ray, mammography and retinal ones. The manual processing of them is very slow, in addition the human factor has a big role too. That is the reason why more and more automated software systems are developed and used saving lots of time and money.
Retinopathy is a tragic effect of diabetes which causes blindness in many cases. With appropriate and accurate screening it can be diagnosed and treated in time avoiding the mentioned side effects. Lesions of retinal vessels – for example change in their tortuosity or width, appearance of microaneurysms – are indicators of some abnormality or disease.
In this thesis I am examining methods, procedures and approaches that support this work. My goal was developing an algorithm for analyzing the retinal vasculature system, and I attempt to create an efficient microaneurysm-detector method. Processing of retinal images is a very important and constantly developing research area. It is clearly shown by the increasing number of new publications.