Efficient Segmentation of Retinal Blood Vessels in The Fundus Images Using Morphological Image Processing

The segmentation of retinal blood vessels is critical in diagnosing and monitoring a variety of retinal ‎disorders, such as hypertension, diabetes, glaucoma, and other ocular and cardiovascular problems. ‎Retinal fundus pictures can aid in the early identification and management of these disorders by ‎providing essential details about the status of the blood vessels in the retina. Manual evaluation of retinal ‎fundus pictures by qualified ophthalmologists or experts takes time and requires competence. To address ‎these issues, an automated system for the purpose of diagnosing and treating eye illnesses, a ‎computerized image processing tool to divide the blood vessels that comprise of the ‎retina. The vessel extraction methodology is carried out in three stages: preprocessing, blood vessel ‎segmentation, and performance metric analysis. The green channel is extracted from the input image ‎during preprocessing, and the contrast is enhanced using CLAHE. After preprocessing, the image ‎undergoes a morphological operation followed by Otsu thresholding to segment the image has pixels that ‎represent vessels and non-vessels, respectively. This method is validated on the DRIVE and HRF ‎datasets, and the performance is measured in terms of accuracy, sensitivity, and specificity‎.