A Multi-Class Retinal Fundus Image Dataset for Deep Learning-Based Ocular Disease Diagnosis

This dataset contains high-resolution retinal fundus images collected from Rajbari Eye Clinic and Specialised Hospital, Bangladesh. It comprises 1,222 original retinal fundus images belonging to 11 disease classes, including Mild Diabetic Retinopathy (67), Moderate Diabetic Retinopathy (196), Severe Diabetic Retinopathy (55), Proliferative Diabetic Retinopathy (17), Dry Age-Related Macular Degeneration (32), Wet Age-Related Macular Degeneration (44), Mild Glaucoma (60), Moderate Glaucoma (136), Severe Glaucoma (235), Advanced/End-stage Glaucoma (335), and No Age-Related Macular Degeneration (45). To address class imbalance and improve model generalization, data augmentation techniques, including horizontal flipping, vertical flipping, rotation, zooming, brightness adjustment, and contrast enhancement, were applied, generating an additional 5,500 augmented images. This dataset is intended for research in artificial intelligence, machine learning, deep learning, transfer learning, explainable artificial intelligence (XAI), computer-aided diagnosis, and automated multi-class retinal disease classification.