Tunisian OCT Dataset for Multi-Disease Retinal Classification

The Retinal-OCT-Clinical-Dataset is a curated collection of clinically acquired retinal Optical Coherence Tomography (OCT) images, developed to support research in automated retinal disease analysis and to serve as a benchmark for deep learning-based classification models.

The dataset includes four diagnostic categories:• Age-related Macular Degeneration (AMD)• Diabetic Macular Edema (DME)• Rhegmatogenous Retinal Detachment (RRD)• Normal (healthy) cases

All images were collected in a clinical setting and annotated by experienced retinal specialists to ensure high diagnostic reliability. The dataset reflects real-world clinical distribution and imaging variability. Importantly, it represents one of the larger OCT datasets focusing on AMD, DME, and RRD from a North African population, contributing to improved demographic diversity in retinal imaging research.

Access Policy: To comply with ethical regulations and patient privacy requirements, the dataset is distributed under controlled access. Researchers may request access for non-commercial scientific purposes under a Data Use Agreement.

Potential applications include:• Training and validation of deep learning models for multi-class retinal disease classification• Benchmarking algorithmic performance across heterogeneous retinal pathologies• Development of segmentation, transfer learning, semi-supervised learning, and self-supervised learning approaches