Glaucoma Disease Detection Using Deep Learning

Abstract

Glaucoma is an eye condition that causes the retina to slowly deteriorate over time. If the disease is detected early enough, its progression can be stopped. Unfortunately, early diagnosis is rare because there are usually no obvious signs in the early stages. Early glaucoma detection is essential because glaucoma with a delayed diagnosis can cause permanent vision loss. It has been demonstrated that computer vision systems can detect glaucoma efficiently and accurately. There are some existing methodologies SVM, KNN, and Random Forest using text datasets and with a low accuracy rate. In this project, we apply deep learning models that can recognize the complex features needed for classification tasks, including microaneurysms, exudate, and retinal hemorrhagic. The project's goal is to suggest a hybrid or innovative method employing CNN that overcomes the limitations of competing techniques and yields more accurate results.