Automated Diagnosis of Retinal Edema from Optical Coherence Tomography Images

Abstract

Retinal edema (RE) is commonest macular syndrome which is caused by formation of cyst segments within intraretinal pathology. RE can cause severe visual impairments or even blindness. Optical coherence tomography (OCT) imaging is the recently introduced eye testing technique that can detect early syndromes of retinal pathology. Many researchers have worked on the automated diagnosis of retinal pathology from fundus images. However, it is very difficult to detect early RE from fundus images and here we propose a discriminant analysis (DA) classifier based fully automated algorithm for the classification of RE using OCT images. We studied 75 OCT scans of 62 patients in which 14 persons had RE and 48 patients were healthy. Our proposed system achieved 100% accuracy for classification of RE patients and 91.86% for healthy subjects.

Keywords: Retinal edema (RE), Linear discriminant analysis (LDA), Optical coherence tomography (OCT) imaging, Fundus photography