Retina image–based optic disc segmentation

Ching-Lin Wang,Jui-Ming Chen,Yi-Wen Hung,Kwong-Chung Tung,Ming-Yuan Hsieh,Meng-Hsiun Tsai,Mao-Hsiang Chan

doi:10.1177/1687814016649298

Abstract

The change of optic disc can be used to diagnose many eye diseases, such as glaucoma, diabetic retinopathy and macular degeneration. Moreover, retinal blood vessel pattern is unique for human beings even for identical twins. It is a highly stable pattern in biometric identification. Since optic disc is the beginning of the optic nerve and main blood vessels in retina, it can be used as a reference point of identification. Therefore, optic disc segmentation is an important technique for developing a human identity recognition system and eye disease diagnostic system. This article hence presents an optic disc segmentation method to extract the optic disc from a retina image. The experimental results show that the optic disc segmentation method can give impressive results in segmenting the optic disc from a retina image.

Highlights

The retina lies on the inner surface of eye, and the inner retina layer is a light-sensitive thin membrane lining in the inner surface of the vascular membrane
This section investigates the performances of the optic disc segmentation (ODS) method by experiments
The black circles represent the segmentation results obtained by the ODS method, the red circles indicate the optic disc region drawn by ophthalmologists and the green closed curves are the optic disc contours provided by the statistics-based method

Summary

Introduction

The retina lies on the inner surface of eye, and the inner retina layer is a light-sensitive thin membrane lining in the inner surface of the vascular membrane. 10,000– 12,000 nerve fibre ganglion cells or axons leave or enter the eyeball through the optic disc, forming the optic nerve in this area, at which the starting points of the blood vessels of retina are located. Locating and segmenting the optic disc involves difficulties because of individual factors, such as blood vessel distribution, protein secretion, image pollution, ocular fundus disease, vascular occlusion and blurred image edges. Chen et al.[6] proposed a statistics-based method to draw the optic disc contour in a retina image. Based on the drawn initial contour, it precisely segments the optic disc further by using the ACM This method is susceptible to retinopathy or eye excretion. The ODS method is indifferent to indistinct object contour which is caused dilating the white object in I by S, the formula can be by excretion around the optic disc

Related works

Experimental results and discussions

Method

Conclusion