Dr. Abdul Khalid – Professor, Dept. of IT, NIET, Greater Noida, India.
Abstract
Diabetic Retinopathy (DR) is a severe eye condition that requires **early detection to prevent vision loss**. This study presents **RetinoNet**, an AI-powered system utilizing **ResNet architecture** to analyze **retinal images** efficiently. By leveraging deep learning techniques, the model provides **accurate, accessible, and rapid DR screening**, reducing the dependence on specialized eye professionals in resource-limited areas.
Keywords
Diabetic Retinopathy, Deep Learning, Retinal Image Analysis, Automated Diagnosis, Medical Imaging, AI in Healthcare, Ophthalmology.
Conclusion
The **ResNet-based RetinoNet model** demonstrates **exceptional accuracy**, reaching **99.5% training accuracy and 99.1% validation accuracy** on the **APTOS 2019 Blindness Detection dataset**. Future improvements will focus on **real-time deployment, mobile applications, and broader dataset integration** for enhanced robustness.
References
Phan, U. et al. (2025). “Progressive Transfer Learning for Multi-Pass Fundus Image Restoration.” arXiv.
Raj, G. M. et al. (2024). “Federated Learning for Diabetic Retinopathy Diagnosis.” arXiv.
Dai, L. et al. (2021). “A Deep Learning System for Detecting Diabetic Retinopathy Across the Disease Spectrum.” Nature Communications.
2 comments