Image Processing Techniques in the Detection of Hemorrhages in Retinal Images (STARE & DRIVE)

There are many different types of diabetic eye problems, but the most common one is called diabetic retinopathy (DR). It is a common cause of blindness around the world. In 2034, the International Diabetes Federation (IDF) says that there will be 552 million people with diabetes, which is a lot. In computer science, new things like advanced image processing, artificial intelligence (AI), and deep learning (DL) have made it easier to find DR before it gets worse. This means that the chances of a person getting better and not losing their eyesight in the future will go up. In this work, the suggested method includes basic preprocessing, the segmentation of hemorrhages with the help of morphological operations, and the connected components labeling. The STARE (Structured Analysis of the Retina) and DRIVE (Digital Retinal Images for Vessel Extraction) datasets were used in this study. To demonstrate the vitality of the chosen methodology, performance measures such as accuracy, specificity, and sensitivity were determined. The accuracy, specificity, and sensitivity of the results of the proposed method are shown to be the highest for images from the STARE(DRIVE) dataset, with values of 85(88) percent, 82(94) percent, and 89(63) percent for the images given.