Diagnosis of plus diseases for the automated screening of retinopathy of prematurity in preterm infants

摘要

The prevalence of Retinopathy of Prematurity (ROP) among preterm infants in the developing countries like India has increased many fold during the past decades. The main cause of this problem is lack of awareness among people, improper diagnostic methods, inter expert variability while screening etc. Treatment of ROP, one of the leading causes of childhood blindness is warranted based on the diagnosis of Plus disease. Comparison of tortuosity of blood vessels with a gold standard image was one of the common techniques used for its screening. Quantitative assessment of the disease includes evaluation of blood vessel tortuosity and vascular dilation in the retinal image. We present a semi-automated computer-based method for the assessment of Plus disease in ROP. This method involves initial preprocessing of the image followed by evaluation of tortuosity and width of retinal blood vessels. Based on the implemented work two image databases, EIARG2 and KIDROP were quantitatively classified as plus, pre-plus or normal case with high accuracy.