SEGMENTATION OF RETINAL VEINS PATTERNS USING MORPHOLOGICAL OPERATORS

Abstract

The purpose of this project is to develop an efficient system based on Artificial Intelligence that detects the distortion with human eye retinal vessels. This project will help doctors to identify the diseases by evaluating retinal veins pattern. In human eye, the retinal vessels structure help to diagnose multiple diseases. The information in this structure help the doctors to predict the complexity of disease in the patients. For this purpose, to evaluate and analyse eye retinal structure manually is a very complex method. Moreover, it is time and money consuming too. In this project, morphological operators will be applied in order to do the segmentation of blood vessels. Firstly, we will obtain eye fundus images and divide it into three channels red, green and blue. The channel preferred for segmentation is usually green channel that carries detail about a human eye structure. Secondly, in order to obtain clear geometry vessel image adaptive threshold 5x5 is applied. Thirdly, image is sharp. Fourthly, 3x3 filter will be applied to it to de noise and soften an image pixel. Fifthly, threshold will be applied to soften images and then morphological operators are used on a grey scale images. A last, the retrieved image will be compared with the healthy structure in order to calculate the difference of structure and to diagnose the possible disease.