Open AccessA low-cost AI- Powered System for Early Detection of Diabetic Retinopathy and Ocular diseases in resource Limited Settings
Author(s) -
Harpreet Vohra,
Mohammad Kamrul Hasan,
Siti Norul Huda Sheikh Abdullah,
Shayla Islam,
Abdul Hadi Abd Rahman,
Shreya N. Bhojake,
Dhruv Bhobal,
Awal Gandhi,
Muskan Garg,
Chandramohan Dhasarathan
Publication year - 2025
Publication title -
ieee access
Language(s) - English
Resource type - Magazines
SCImago Journal Rank - 0.587
H-Index - 127
eISSN - 2169-3536
DOI - 10.1109/access.2025.3572471
Subject(s) - aerospace , bioengineering , communication, networking and broadcast technologies , components, circuits, devices and systems , computing and processing , engineered materials, dielectrics and plasmas , engineering profession , fields, waves and electromagnetics , general topics for engineers , geoscience , nuclear engineering , photonics and electrooptics , power, energy and industry applications , robotics and control systems , signal processing and analysis , transportation
Eye Health is a vital yet often neglected aspect of healthcare, especially in distant and deprived regions that are less accessible to specialized diagnostic tools and professionals. This paper introduces an intelligent diabetic retinopathy detection system to enhance affordability, accessibility, and early diagnosis. The system features a cost-effective imaging setup utilizing a 28D lens and a camera to capture high-quality fundus images. These images are analyzed using advanced Convolutional Neural Networks to determine the cases of people suffering from eye issues: Age-related vascular degeneration (AMD), Diabetic retinopathy (DR), and Glaucoma. A user-friendly web application hosted on AWS facilitates seamless image upload, automated diagnosis, and report generation. Achieving an accuracy of 96.5%, the system ensures reliable diagnostics, enabling timely medical intervention. By integrating AI-driven analysis and cloud-based accessibility, this solution aims to transform eye care in resource-limited settings. Performance evaluation against existing methods highlights its superiority, with an F1 score of 98.2 % for DR, 92.5% for Glaucoma, and 98.8% for AMD, outperforming many previous approaches that either focused on single–disease detection or relied on costly hardware.