Abstract
Glaucoma is a group of diseases that damage the eye’s optic nerve. If not treated, glaucoma can lead to vision loss or permanent blindness. Glaucoma is very problematic because often there are no symptoms in the early stages. Once symptoms appear, it could be too late to prevent blindness. There exist several methods to detect glaucoma such as tonometry, which examines the inner eye pressure, ophthalmoscopy, which examines the shape and color of the optic nerve, and perimetry, which tests for the complete field of vision. However, these methods do not allow to detect all types of glaucoma as the conditions vary from one person to another. Cup-to-disc ratio is a visual inspection method that can detect glaucoma. Image processing techniques such as binarization and superpixel classification are used in this project to calculate the cup-to-disc ratio and use it to determine if an eye has glaucoma. Throughout the project, various details are needed for analyzing fundus images to detect glaucoma with image processing techniques. It is shown that the error margin is sensitive to many factors, including the color and quality of the image and the presence of blood vessels. The results are not 100% accurate but with acquirement of better photographs and using more advanced technologies such as 3D imaging, the proposed methods can be improved in terms of the computation speed and accuracy.