Glaucoma Classification Using Gossip Learning

Abstract

Glaucoma is the leading cause of irreversible blindness worldwide and poses significant diagnostic challenges due to its reliance on subjective evaluation. However, recent advances in computer vision and deeplearning have demonstrated the potential for automated assessment. In this paper, we survey recent studies on Gossip Learning-based glaucoma diagnosis using fundus, optical coherence tomography, and visual field images, with a particular emphasis on deep learning-based methods. We provide an updated taxonomy that organizes methods into architectural paradigms and includes links to available source code to enhance the reproducibility of the methods. Through rigorous benchmarking on widely-used public datasets, we reveal performance gaps in generalizability, uncertainty estimation, and multimodal integration. Additionally, our survey curates key datasets while highlighting limitations suchas scale,labelinginconsistencies,andbias.We outline openresearchchallenges anddetailpromising directions for future studies. This survey is expected to be useful for both Gossip Learning researchers seeking to translate advances into practice and ophthalmologists aiming to improve clinical workflows and diagnosis using the latest Gossip Learning outcomes.

Introduction

Glaucoma, a major cause of irreversible blindness, requires early detection and accurate classification for effective treatment. Traditional machine learning approaches for glaucoma diagnosis face challenges related to patient data privacy and the need for large datasets. Gossip learning (a form of federated learning) offers a solution by enabling collaborative training across multiple institutions without sharing sensitive patient data, thus preserving privacy while improving model robustness.

This study proposes gossip pVIT, a novel approach combining Visual Information Technology (VIT) for feature extraction with gossip learning to classify glaucoma stages using decentralized datasets. The method improves classification accuracy and respects data confidentiality.

Key objectives include:

• Designing a gossip learning architecture for collaborative multi-institutional training without data sharing.

• Enhancing classification accuracy via advanced VIT methods.

• Validating the model’s performance across diverse datasets.

• Addressing privacy concerns using encryption and differential privacy.

• Promoting secure collaborative research among healthcare institutions.

The methodology involves:

• Collecting and annotating retinal images from multiple institutions.

• Preprocessing images for robustness.

• Local model training on institution-specific data.

• Aggregating model updates at a central server via gossip averaging without accessing raw data.

• Iterative model refinement and hyperparameter tuning.

• Ensuring privacy through encryption and differential privacy techniques.

• Deploying the system in clinical workflows and fostering ongoing collaboration.

This approach aims to improve glaucoma classification accuracy, safeguard patient privacy, and facilitate collaborative healthcare research through decentralized AI training.

Conclusion

In conclusion, this paperpresentsthegossip framework as a pioneering approach to glaucoma classification usingfederatedlearning. Byprioritizingpatientprivacy and fostering collaborative research, aims to advance diagnostic capabilities in ophthalmology. The promising preliminary results suggest that federated learning can effectively harness distributed datasets to improve model performance, ultimately contributing to better patient outcomes in glaucoma management. Future work will focus on refining the model, expanding collaborative networks, and exploring additionalapplications of gossiplearning in healthcare.

Copyright

Copyright © 2025 T. C. Kalaiselvi, R. Swetha, G. Pradeep, N. Narasimman, G. Vidubala. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.