Automatic Diagnosis of Glaucoma using Image Processing Technique

Abstract

Glaucoma is a human eye condition which will affect the optic nerve present in the retina. This condition occurs due to the abnormal ocular pressure in human eye. If it is not diagnosed and treated well in advance, it may lead to blindness. This is the major problem of elderly people all over the world. The best way to avoid vision loss due to glaucoma is to detect the disease at the early stage and treat it as soon as possible. These are the keys to prevent blindness. As vision is an important organ in human body it is advisable to keep it healthy. The optic cup in the retina will be pulled in towards the optic nerve away from the optic disc. At one point, the cup will be detached from the retina, causing blindness. So if one can monitor by measuring the optic disc to cup ratio, the progression of glaucoma can be diagnosed earlier. The proposed method detects the optic disc and cup using thresholding method. Direct least square fitting algorithm is used here to fit the ellipse in order to calculate the cup height and disc height. Then the ratio is calculated. If the calculated ratio is above the threshold value, it is considered as glaucoma affected eye otherwise not. The CDR is calculated using the formula VDH/VCH (Vertical Disc Height to the Vertical Cup Height). Thus, the proposed method helps to automatically detect the glaucoma disease with better sensitivity and specificity.