A Web-based Platform for Automated Diabetic Retinopathy Screening

Abstract

Diabetic patients are encouraged to undergo frequent retinal examination because of diabetic retinopathy (DR): the most common cause of blindness in working population. As this population is generally too large for healthcare systems, exam reviewing must be optimized. Therefore, a web-based application is proposed. It features user interfaces for healthcare professionals, including ophthalmologists, and an automated DR detection module (by means of image processing and diverse computational intelligence techniques) that allows to relieve their work load. Overall automated system performance reaches 91.9% and 65.2% in terms of sensitivity and specificity, respectively.