AI-Based Primary Diagnosis of Ocular Health via Smartphone

Traditional diagnostic methods such as Fundus cameras and slit-lamps that are used for diagnosing various ocular diseases are often limited by their cost and accessibility which hinders timely intervention especially in rural areas and indigent societies with limited resources. While the current research is mostly focused on classification of single diseases or expensive setups, the proposed work aims to address these limitations by developing a cost-effective and user-friendly system for primary diagnosis of various ocular diseases, including Diabetic Retinopathy, Age-Related Macular Degeneration and Glaucoma. The proposed work utilizes a smartphone, cloud-based deep learning model, and a 3D printed multi-lens holder for fundus imaging. This research leverages open-source datasets for model training and testing, added to new sets of images that were collected throughout this work for validation. The new set tested the quality of the captured images in actual setup and verified accurate classification. Based on the results, the system has been successfully implemented and the overall classification accuracy of the multi-classification ResNet-50 model is 74%. Preliminary results demonstrate the potential of this approach for multi-classification of ocular diseases, marking a significant advancement in portable ocular health primary diagnostics.