Early Glaucoma Detection using Deep Learning with Multiple Datasets of Fundus Images

TL;DR

This paper introduces a deep learning pipeline using EfficientNet-B0 trained on multiple fundus image datasets to improve early glaucoma detection, demonstrating strong generalization and high AUC-ROC performance.

Contribution

It presents a novel multi-dataset training approach with minimal preprocessing for effective glaucoma detection using deep learning.

Findings

Minimal preprocessing outperforms complex enhancements in AUC-ROC

Model generalizes well across unseen datasets

Deep learning pipeline achieves high discriminative accuracy

Abstract

Glaucoma is a leading cause of irreversible blindness, but early detection can significantly improve treatment outcomes. Traditional diagnostic methods are often invasive and require specialized equipment. In this work, we present a deep learning pipeline using the EfficientNet-B0 architecture for glaucoma detection from retinal fundus images. Unlike prior studies that rely on single datasets, we sequentially train and fine-tune our model across ACRIMA, ORIGA, and RIM-ONE datasets to enhance generalization. Our experiments show that minimal preprocessing yields higher AUC-ROC compared to more complex enhancements, and our model demonstrates strong discriminative performance on unseen datasets. The proposed pipeline offers a reproducible and scalable approach to early glaucoma detection, supporting its potential clinical utility.