Next-Gen Ophthalmic AI: Early Eye Disease Detection with ResNet-18

Classifying eye diseases is an important issue for countries around the world. The goal is to use advanced computer methods, created from research, to improve old models. This will help us accurately identify different eye diseases by looking at photos. This paper used machine learning and computer vision techniques to sort images into four categories of eye problems: cataract, diabetic retinopathy, glaucoma, and healthy eyes. In this study, we taught the ResNet-18 model again using a learning rate of 5x10^-5 for the ResNet-18 part and 8x10^-4 for the extra dense layers. Our model worked really well, with an overall accuracy of 86%. The model did a great job in identifying diabetic retinopathy cataract, showing strong accuracy and effectiveness in its results. However, other groups of people with glaucoma had much lower scores on these measures. These results strongly show that the model could help find, diagnose, monitor, and treat eye diseases early. This will be very helpful for people and their medical care. This will help society by improving the overall quality of medical care. Doctors and nurses everywhere are using these results as a guide to help provide good care at a fair price in the healthcare system.