Computer Aided Optical Disc Detection in Fundus Images: A Review

Abstract

Colour fundus images of the human retina are increasingly used in the diagnosis and treatment of several eye related pathologies and in ailments such as arteriosclerosis, diabetes, and hypertension. The effectiveness of treatment for many eye related diseases lies in the early detection through regular screenings. But, screening a large number of patients is a significant problem faced by medical practitioners in populous developing countries like India. In addition, there are extensive influences of human blunders and subjectivity on the consequences of assessment by a human master. This opens up the probability of applying propelled image processing techniques in fundus images to support and improve determination in different ways. Thus, automatic detection of pathologies and computer aided analysis in retinal images play a important role in modern diagnostic procedures and screening systems. Reliable and robust extraction of retinal features like optic disc, macula and vasculature is an essential for ensuing retinal image analysis and handling since these are the overwhelming and most stable structures showing up in the retina. However, automatic segmentation of retinal images is a complicated affair since retinal images are often poorly contrasted noisy, and there are a wide variations in orientation. This paper presents a review on algorithms or methods to detect optic disc automatically.