Detection of Glaucoma on Fundus Images Using Deep Learning

Abstract

Detecting glaucoma at an early stage is crucial in preventing irreversible vision loss. Deep learning (DL), especially models like Convolutional Neural Networks (CNNs) and Vision Transformers (ViT), has demonstrated significant potential in medical image analysis. This study introduces a glaucoma detection system that leverages deep learning for fast, accurate, and reliable diagnosis using retinal fundus images. Three models are used in this system: ResNet50, Vision Transformer (ViT), and a hybrid model combining both.