Diagnosis of Choroidal Neovascularization and Detection of Drusens for Diagnosis of AMD Using Retinal OCT Images

Abstract

Age Related Macular Degeneration (AMD) is a type of progressive eye disease which occurs in patients who are above 50 years in age. AMD encompasses many diseases such as Choroidal Neo-Vascularization (CNV), diabetic retinopathy, glaucoma etc. it is caused when cholesterol deposits in the retinal layers or with the growth of abnormal blood vessels in the retina. This causes damage to the photoreceptors in the eye resulting in acute drop in central vision. In Pakistan around 15 million people suffer from AMD making up 10% of the total population. Use of Optical Coherence Tomography has been introduced in Pakistan for the cross sectional imaging of retinal layers. The proposed system makes use of software engineering techniques for the iterative elicitation of requirement gathered from AFIO. Using modular approach the system is broken down into chunks for efficient requirement elicitation & specification. It devises a diagnostic algorithm for the detection of CNV in the retinal layers and detects the presence of drusens in the retinal pigment epithelial (RPE) layer. Achieving an overall accuracy of 94% for CNV detection and 85% for drusen detection, this system is an example of amalgamation of software engineering techniques in telemedicine. It is a step towards making Pakistan capable of advancement in telemedicine.