Computer-Aided Assessment of Longitudinal Fundus Photos for Screening Diabetic Retinopathy

Abstract

Diabetic retinopathy (DR) is a complications of diabetes mellitus, which progressively damages small retinal blood vessels and result in vision loss if not treated and controlled timely. Because of an increase in the risk of vision loss with the duration of diabetes and the latency between DR progression and early symptoms, diabetic patients require periodic screening. The required regular screening by a trained clinician, based on fundus photos, is time consuming, subjective, and resource demanding. Furthermore, the current practice does not scale well with the global rise in the diabetic population. Computer-aided screening offers a solution to this problem. This thesis presents several building blocks for automated analysis of a series of fundus images for DR.