Glaucoma Detection Using Deep Learning and Image Processing

Glaucoma is a progressive eye condition that can silently damage the optic nerve and lead to irreversible blindness if not identified early. Because early symptoms are subtle, effective screening tools are essential. In this study, we present a practical diagnostic framework that integrates classical image enhancement with a lightweight convolutional neural network (CNN) for retinal fundus analysis. The system improves image clarity, isolates the optic disc and cup, and calculates the cup-to-disc ratio (CDR), a clinically recognized marker of glaucoma. A CNN classifier then categorizes images into normal, suspect, or glaucomatous cases. To improve accessibility, a Tkinter-based interface allows users to upload images and view diagnostic outputs in real time. Experimental evaluation demonstrates strong accuracy, sensitivity, and specificity, highlighting the system’s potential as a cost- effective solution for early glaucoma screening and clinical support.