Diabetic Retinopathy Detection Using Smartphone Fundoscopy: Advancing Healthcare Equity through Intelligent Edge Computing

Diabetic Retinopathy (DR) is a serious consequence of diabetes that can result in the loss of vision if not identified and addressed early. Conventional screening procedures encounter substantial obstacles, particularly in distant and underserved regions, as result of geographical, economic, and resource limitations. This research introduces a new method that takes advantage of the widespread use of cellphones and the capabilities of intelligent edge computing to make DR screening accessible to everyone. Retinal images of excellent quality are obtained and analyzed using the EfficientNetB0 model, a cutting-edge convolutional neural network, through smartphone fundoscopy with a 20D lens. The model, which has been optimized for detecting diabetic retinopathy (DR), exhibits exceptional accuracy, offering a cost-efficient and easily accessible approach for early identification of DR. This method not only improves the availability and fairness of healthcare but also lessens the strain on healthcare systems and empowers individuals and communities to actively engage in managing their health. The implementation of this method through a user-friendly mobile application showcases considerable potential in revolutionizing healthcare provision and tackling crucial health inequalities on a worldwide scale.