AEye Doctor: An Automated Diagnosis System for Ophthalmological Diseases

By Tienlan Sun
Intermediate Category (Grades 9-10)
Innovation | Big Data / AI, Biology

Early detection of ophthalmological diseases is a key factor in the prevention and treatment of vision impairment and blindness. However, due to the global shortage of health professionals and expensive diagnosis equipment, this is unachievable through traditional diagnosis methods. This project presents a novel automated diagnosis system, AEye Doctor, for the population screening of ophthalmological diseases.

Using retinal images as input, it diagnoses six common ophthalmological diseases including age-related macular degeneration, cataracts, diabetic retinopathy, glaucoma, hypertensive retinopathy and pathological myopia. By training a deep learning algorithm on small datasets of ethnically diverse retinal images, high specificity, sensitivity and area under receiver operating characteristic curves (AUROC) were achieved. The system is accessible on all devices through web browsers. In addition to providing the diagnosis and probability of each disease, AEye Doctor will also display adjustable saliency heatmaps and a personalized artificial intelligence chatbot. The saliency heatmaps will improve ophthalmologist diagnosis accuracy by highlighting the areas of the retinal images that contributed the most to the final diagnosis. The AI chatbot will provide easily understood information to the patient depending on individual diagnosis conditions.

By allowing the patient to become informed prior to meetings with medical professionals, efficiency can be further increased. The improvement in treatment efficiency will yield improved patient outcomes, as greater numbers of patients will receive early diagnosis. With these properties, AEye Doctor has the potential to combat the global lack of ophthalmologists as a tool for the initial screening of ophthalmological diseases without supervision.

Related Project


Smart DeTech