Integrating Fundus and OCT Imaging for Glaucoma Detection: An AI Perspective

Abstract

Glaucoma is one of the leading causes of irreversible blindness worldwide, and early detection is essential to prevent permanent loss of vision. Recent advances in artificial intelligence (AI) have enabled automated analysis of ophthalmic imaging modalities such as retinal fundus photography and optical coherence tomography (OCT). These technologies allow the extraction of structural biomarkers including optic disc and optic cup boundaries, retinal nerve fiber layer thickness, and vessel density for accurate diagnosis. This paper presents a comprehensive survey of recent AI-based glaucoma detection approaches that utilize fundus imaging, OCT imaging, and multimodal diagnostic frameworks. The study analyzes twenty state-of-the-art models and compares their datasets, methodologies, reported accuracies, and limitations. The findings show that deep learning architectures such as CNNs, U-Net variants, and transformer-based models significantly improve detection accuracy compared with traditional machine learning techniques. The analysis indicates that the reported diagnostic accuracies range between 87% and 99.5%, reflecting steady progress in the effectiveness of AIassisted glaucoma detection. The survey also highlights existing challenges and future research opportunities for developing scalable, interpretable, and clinically deployable glaucoma screening systems.