Functional Localization Enforced Deep Anomaly Detection Using Fundus Images

TL;DR

This study evaluates a Vision Transformer classifier for retinal disease detection from fundus images, demonstrating its robustness across datasets and introducing a GANomaly-based anomaly detector with explainability for clinical use.

Contribution

It systematically assesses ViT performance with augmentation strategies across multiple datasets and proposes a GANomaly-based anomaly detector with explainability.

Findings

ViT achieved 0.789 to 0.843 accuracy across datasets.

ViT with geometric augmentation achieved 0.91 AUC on Papila dataset.

GANomaly-based detector achieved 0.76 AUC with explainability.

Abstract

Reliable detection of retinal diseases from fundus images is challenged by the variability in imaging quality, subtle early-stage manifestations, and domain shift across datasets. In this study, we systematically evaluated a Vision Transformer (ViT) classifier under multiple augmentation and enhancement strategies across several heterogeneous public datasets, as well as the AEyeDB dataset, a high-quality fundus dataset created in-house and made available for the research community. The ViT demonstrated consistently strong performance, with accuracies ranging from 0.789 to 0.843 across datasets and diseases. Diabetic retinopathy and age-related macular degeneration were detected reliably, whereas glaucoma remained the most frequently misclassified disease. Geometric and color augmentations provided the most stable improvements, while histogram equalization benefited datasets dominated by structural subtlety. Laplacian enhancement reduced performance across different settings. On the Papila dataset, the ViT with geometric augmentation achieved an AUC of 0.91, outperforming previously reported convolutional ensemble baselines (AUC of 0.87), underscoring the advantages of transformer architectures and multi-dataset training. To complement the classifier, we developed a GANomaly-based anomaly detector, achieving an AUC of 0.76 while providing inherent reconstruction-based explainability and robust generalization to unseen data. Probabilistic calibration using GUESS enabled threshold-independent decision support for future clinical implementation.