Analysis of Visual Field Reports for Diagnosis of Glaucoma

Abstract

Glaucoma is the second leading cause of blindness in the world, according to the World Health Organization, the first being cataracts.Everyone is at risk for glaucoma from babies to senior citizens. Glaucoma is an eye disease that is characterized by a particular pattern of progressive damage to the optic nerve that generally begins with a subtle loss of side vision (peripheral vision). Elevated pressure in the eye is the main factor leading to glaucomatous damage to the eye (optic) nerve. If not diagnosed and treated, it can progress to loss of central vision and blindness causing an irreversible damage.Glaucoma usually causes no symptoms early in its course, at which time it can only be diagnosed by regular eye examinations. Visual fields are used to diagnose the presence of glaucoma and monitor its progression .The blind spots created in the patient’s eye by glaucoma can be seen on the visual field map. This type of test is known as Standard Automated Perimetry (SAP). There are various types of SAP, but the most commonly used is Humphrey.In this thesis we use image processing and decision support techniques to automate the analysis of visual field reports in order to aid ophthalmologists. The main focus is extract, locate and score quantitatively the glaucomatous damage in each hemisphere so that the extent of damage may be known. The pattern deviation (PD) plot printed in the visual field test report contains the significantly depressed points referred to as probability key symbols (PKS); whose shape, location in ISNT regions and count give us information about the damage caused by glaucoma to the optic nerve. Our thesis focuses on the extraction, location region, count and score of these PKSs in the upper and lower hemispheres of the PD plot hence automating the scoring of the analysis of visual field tests