An Artificial Intelligence-Based Smart System for Early Glaucoma Recognition Using OCT Images

An Artificial Intelligence-Based Smart System for Early Glaucoma Recognition Using OCT Images

Law Kumar Singh (Department of Computer Science and Engineering, School of Engineering and Technology, Sharda University, Knowledge Park III, Greater Noida, India & Department of Computer Science and Engineering, Hindustan College of Science and Technology, Mathura, India), Pooja (Department of Computer Science and Engineering, School of Engineering and Technology, Sharda University, Knowledge Park III, Greater Noida, India), Hitendra Garg (Department of Computer Engineering and Applications, GLA University, Mathura, India) and Munish Khanna (Department of Computer Science and Engineering, Hindustan College of Science and Technology, Mathura, India)
Copyright: © 2021 |Pages: 28
DOI: 10.4018/IJEHMC.20210701.oa3
Article PDF Download
Open access articles are freely available for download

Abstract

Glaucoma is a progressive and constant eye disease that leads to a deficiency of peripheral vision and, at last, leads to irrevocable loss of vision. Detection and identification of glaucoma are essential for earlier treatment and to reduce vision loss. This motivates us to present a study on intelligent diagnosis system based on machine learning algorithm(s) for glaucoma identification using three-dimensional optical coherence tomography (OCT) data. This experimental work is attempted on 70 glaucomatous and 70 healthy eyes from combination of public (Mendeley) dataset and private dataset. Forty-five vital features were extracted using two approaches from the OCT images. K-nearest neighbor (KNN), linear discriminant analysis (LDA), decision tree, random forest, support vector machine (SVM) were applied for the categorization of OCT images among the glaucomatous and non-glaucomatous class. The largest AUC is achieved by KNN (0.97). The accuracy is obtained on fivefold cross-validation techniques. This study will facilitate to reach high standards in glaucoma diagnosis.
Article Preview
Top

1. Introduction

Glaucoma is the second most crucial optic 'eye' disease in this world. As per sources in 2010, approximately 60 million populations cross-sectional to be disease infected, and this count is increasing above 20 million in the period 2020. It does no repairable damage to the central part of the optic nerves that can lead to making the person blind. Hence, sensing Glaucoma during the initial stage is very imperative. Generally, doctors focus on the area of the optic disc & optic cup and find the edges though optic nerve examination. They assure the glaucoma presence they identify the increased size of the optic cup. One of the essential features is to identify the ratio of the height of the Cup to the disc; this is the crucial indicator for identifying Glaucoma. Among the patients, if the Cup-to-Disc ratio (CDR) value is at least 0.5, it may be considered as the glaucomatous eye.

Human eye mainly has three layers. The outer layer: Sclera, which is used to protect the eyeball; Second layer: Choroid and the innermost layer: Retina.”Retina” is liable in vision because of the presence of photoreceptors. Researchers have invented many techniques which are used for detecting retinal disorder. These techniques include fundus photography, fluorescein angiography, and the OCT.

Figure 1.

OCT image with macular edema

IJEHMC.20210701.oa3.f01

1.1 Optical Coherence Tomography

OCT stands for optical coherence tomography, which is having the property that it is a noninvasive diagnostic technique used for manufacturing or producing the view of the retina in a cross-sectional way (Burgansky-Eliash et al., 2005).OCT imaging technology uses OCT cameras, which have low coherence interferometers in which the low coherence visible light is always allowed to penetrate the entire human retina and it reflects interferometer producing a cross-sectional image of the retina. Low coherence of lightning is used for producing images of resolutions—the basic principles of the OCT imaging technique on the (Michelson type interferometer). A considerable benefit of the OCT imaging system (Fig. 1) is always that it can detect retinal disorders earlier than other techniques.

It has been observed that the majority of the work, to date, accomplished by the researcher's fraternity is performed on fundus images for glaucoma detection. However, we have used OCT images as there are multiple advantages of OCT images over fundus images. OCT has higher sensitivity than the fundus image for the detection of early Glaucoma. It is a non-invasive technology and has a higher ability to detect small changes in the subretinal layer than the fundus image. This image picks up the earliest signs of diseases. OCT evaluates disorders of the optic nerve and changes to the fibers of the optic nerve. It detects changes caused by Glaucoma.

Moreover OCT(Khalil et al., 2014) images also have higher resolution.OCT provides topographical information about retinal ganglion cells and Retinal Nerve Fiber Layer (RNFL) abnormalities.OCT images allow for direct visual analysis of the scans, much the way MRI scans are analyzed, rather than depending entirely upon computer-driven summary statistics. It can produce real-time a cross-sectional image of an object, i.e. a two-dimensional image in the space.OCT images give more accuracy as compared to the fundus images. Analysis of OCT is done by the degradation of a separate layer, which increases the efficiency of an image.OCT provides detailed structural information on clinical abnormalities. This method has been extensively applied to the setting of ophthalmology; this method is popular due to its ability to perform high-resolution cross-sectional imaging and diagnosing the changes in the structure of the eye during the progression of this disease.

Complete Article List

Search this Journal:
Reset
Volume 13: 2 Issues (2022): 1 Released, 1 Forthcoming
Volume 12: 6 Issues (2021)
Volume 11: 4 Issues (2020)
Volume 10: 4 Issues (2019)
Volume 9: 4 Issues (2018)
Volume 8: 4 Issues (2017)
Volume 7: 4 Issues (2016)
Volume 6: 4 Issues (2015)
Volume 5: 4 Issues (2014)
Volume 4: 4 Issues (2013)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing