Automated diabetic retinopathy grading and screening using deep learning.

AI / informaticsInternational journal of retina and vitreous· Jun 2026

Automated diabetic retinopathy grading and screening using deep learning.

Ahmed M Nashaat Ali Rady, Mohamed A Azim, Ahmed AbdelMoety, Mohamed Abdelmoaty Ahmed

Read paper

RetinaAI / informaticsDiagnosisQ1IF 3.6

Deep learning models show high accuracy for binary diabetic retinopathy screening.

Automated diabetic retinopathy grading and screening using deep learning.

Ahmed M Nashaat Ali Rady … Mohamed Abdelmoaty Ahmed

International journal of retina and vitreous · Jun 2026 · Q1 in Ophthalmology

Purpose

To develop and benchmark a unified Deep Learning (Deep Learning) pipeline for automated detection and five-level grading of Diabetic Retinopathy (Diabetic Retinopathy), and to derive a high-performance binary screening endpoint (Diabetic Retinopathy vs No Diabetic Retinopathy) suitable for scalable use in resource-limited settings.

Methods

A publicly available five-class Diabetic Retinopathy fundus dataset of 3,500 color photographs graded using the International Clinical Diabetic Retinopathy (International Clinical Diabetic Retinopathy) scale was used.

n = 3,500 color photographs

AUROC ≈0.982-0.990

Results

AUROC ≈0.982-0.990

This indicates the models' high accuracy in detecting diabetic retinopathy

n = 3,500 color photographs

More results

For five-class grading, VGG16 achieved the highest accuracy (0.7686) and weighted Jaccard index (0.6572), while EfficientNetV2B2 provided the best balanced accuracy (0.6128) and macro Area Under the Receiver Operating Characteristic Curve (AUROC 0.9158).

Misclassifications were concentrated between adjacent severity levels.

“

Conclusion · 1 of 3

The proposed Deep Learning pipeline delivers robust multiclass Diabetic Retinopathy grading and highly discriminative binary screening using widely available Convolutional Neural Network backbones.

Conclusion · 2 of 3

Operating-point calibration enables sensitivity-oriented triage or specificity-oriented confirmation, supporting teleophthalmology and task-shifted Diabetic Retinopathy screening programs to expand coverage and reduce preventable vision loss.

Conclusion · 3 of 3

CLINICAL TRIAL NUMBER: Not applicable.

Automated diabetic retinopathy grading and screening using deep learning.

Automated diabetic retinopathy grading and screening using deep learning.

Related papers