Color analysis in retinography: Glaucoma image detection

Abstract

Glaucoma is a leading cause of permanent blindness. Glaucoma has an initial phase difficult to diagnose, so different measures are used to evaluate it. One of them is a retinography exam which is used to determinate the ratio between the cup and the disc areas, and vein shifting in disc perimeter. A system that allows different measures and calculate them automatically, could be useful to medical community. In this paper it shows a study about papilla and cup segmentation, their ratio and vein shifting ratio and their classification using neural networks. Our system is able to classify the glaucoma automatically with a sensitivity and specifity of 96,77% and 68,2% respectively. © Springer International Publishing Switzerland 2014.