Design and implementation of fully automated systems to aid in retinopathy of prematurity screening and diagnosis

Abstract

The retina is the inner light-sensitive layer of the eye, which is responsible for human vision. Retinal image analysis has great significance in identifying the changes in the retinal blood vessel structure for detecting the presence of various retinal diseases. Retinopathy of Prematurity (ROP) is a retinal disease affecting the retina of low birth-weight preterm infants and is a leading cause of childhood blindness throughout the world. Preterm birth spurs the growth of abnormal blood vessels which bleed and scar the retina, increasing the risk of retinal detachment and vision loss in infants. The loss of vision could be prevented if the disease is identified and treated at the right stage with respect to disease progression. A vital sign of treatment-requiring ROP or threshold ROP is the presence of plus disease. Plus disease is characterized by abnormal twisting, widening and branching of the retinal blood vessels in the posterior region of the retina, and it requires specially trained ROP experts to detect the disease based on these symptoms.