Automated segmentation of RPE layer for the detection of age macular degeneration using OCT images

Abstract

Macular retinal anomaly known as Drusen present in the eyes infected with Age Macular Degeneration (AMD) can be visualized with optical coherence tomography (OCT). This paper presents an algorithm to automatically segment retinal pigment epithelium (RPE) layer of an eye using spectral domain optical coherence tomography (SD-OCT) images for the detection of drusen. Segmentation framework is based on intensity level thresholding of high reflective pixels around retinal nerve fiber layer (RNFL). A dataset of 50 different B-scan is taken with having 25 AMD and 25 heathy images. Bottom layer of the retina known as RPE layer is extracted and analyzed by fitting a second order polynomial curve. The difference between both the curvess decide the class of the image. Results show an accurate detection of drusen with 96 % accuracy.