A Systematic Review on Deep Learning Approaches for Multi-Class Ocular Diseases Classification using Fundus Images
Description
This is the data associated with the systematic review of eye disease classification using fundus images using deep learning architectures. The review is conducted on studies published between 2015 and February 2026 and follows the PRISMA guidelines. The study categorizes eye disease classification research using fundus images into CNN-based, Hybrid, and Transformer-based approaches. In addition, the commonly used public datasets in these studies and the preprocessing steps are described. Studies were identified through five major databases, and their methodologies, evaluation metrics, and outcomes were systematically analyzed. The main challenges for eye disease classification using fundus images are highlighted, and future research directions are proposed to enhance generalizability, interpretability, and data efficiency.
Files
Institutions
- Manipal Academy of Higher EducationKarnataka, Manipal