Paper Contents
Abstract
This project explores the crucial field of medical imaging on Diabetic Retinopathy(DR), with an emphasis on image quality enhancement. This project focus on enhancing the fundus images which make correct grading easy. Enhancing the raw fundus images and transforming them into high-quality is essential for predicting the grade level in DR diagnosis. For enhancing the fundus images we use a functions and techniques from python library called OpenCV. The trained model AlexNet is used to classify retinal images into diabetic retinopathy severity levels. Leveraging the sequential connectivity of AlexNet, it enables the capture of intricate details and features within retinal images, enhancing subtle abnormalities associated with diabetic retinopathy (DR) such as microaneurysms, hemorrhages, and exudates. Experimentation is performed on diabetic retinopathy dataset, which has 6,040 retinal images with a five classes of severity levels i.emild, severe, moderate, no_dr, proliferate_dr. The mentioned techniques yield an accuracy of 92%, for the identification of severity levels in DR images and it provides the report of guiding about the treatment to be taken.
Copyright
Copyright © 2024 Devadi Ganesh. This is an open access article distributed under the Creative Commons Attribution License.
