AN AUTOMATED AND ROBUST IMAGE PROCESSING ALGORITHM FOR GLAUCOMA DIAGNOSIS FROM FUNDUS IMAGES USING NOVEL BLOOD VESSEL TRACKING AND BEND POINT DETECTION

Description

1.0 Introduction

This research is on an automated and robust image processing algorithm for glaucoma diagnosis from fundus images using novel blood vessel tracking and bend point detection. Glaucoma is neurodegenerative eye disease that impairs the optic nerve which is a vital part of the eyes. It is characterized by structural abnormalities of the optic nerve head (ONH) which enlarges or deepens the optic cup leading to the loss of peripheral vision in the vision field (VF). Also, there is an elevated pressure in the eye know as the intra-ocular pressure (IOP) . This elevated pressure in the eyes can lead to an irreversible blindness. Thus to slow down the progression of the pressure in the eye, timely diagnosis of glaucoma is very important. Glaucoma diagnosis according to ( Kwade and Bairagi, 2016) is made using different source of information on the risk factors of glaucoma like progressive IOP; and different testing methods using different technical tools to examine the pupil, IOP, visual field, and the ONH.

In this study, image processing algorithm will be applied to detect glaucoma from fundus images. The image processing technique allows for the identification of the optic nerve’s shape and colour which is then processed for detecting glaucoma. The image processing  algorithm will analyze the optic disc and optic cup which will be determined using blood vessel tracking and bend point determination.

There are about 3 categories of glaucoma – primary glaucoma, secondary glaucoma and congenital glaucoma (Glaucoma Research Foundation, 2016).

2.0       Literature Review

Several image processing techniques have been used the detect glaucoma from fundus images. Cheng et al. (2015) proposed the use of CDR as a parameter for deteting glaucoma from fundus images. Here, the optic cup and optic disc were segmented using center surround statistics and histogram which classified a fundus image (pixel) as either super-pixel or not super-pixel. Aquino et al., (2010) proposed the use of a template based algorithm  for  optic nerve head segmentation from fundus images. The optic nerve head is then analyzed using morphological arithmetic and edge  detection. Yuji et al., (2018 ) proposed the use of an improved automatic  method to detect glaucoma. Here, the optic cup was segmented based on blood vessels and bend detection. This was determined by tracking the blood vessels from the disc edge to the primary cup edge. These works and man more have indicated the performance of image processing algorithm to detect glaucoma. This study will contribute to already existing literature on the use of image processing algorithm in detecting glaucoma from fundus images.

3.0 Materials and methods

In this section, the algorithm used for the automatic and robust detection of glaucoma from fundus images by tracking blood vessels and determining bends. From the fundus image, the optic cup, optic disc and blood vessels can be identified. A digital colour fundus images consist of red, blue and green colour. These features of the Fundus image is observed in figure 1.