Determination of Early Onset Glaucoma Using OCT Image

Introduction

Axonal degeneration in retinal ganglion cells (RGC) is the hallmark of glaucoma, a heterogeneous group of degenerative neural disorders. Approximately 111.8 million people are projected to have glaucoma in 2040, making it the leading cause of irreversible blindness worldwide.When the optic nerve is damaged, blind spots form in your visual field. This damage is caused by increased pressure inside the eye, for reasons that medical professionals aren't fully aware of.If you have glaucoma, you will need to lower your eye pressure (intraocular pressure). Patients with glaucoma commonly report experiencing the following visual symptoms as a result of their glaucoma: Blurry vision, glare, and needing more light.This disease leads to vision loss and increased pressure within the eye, as well as chronic and progressive optic neurodegenerative changes.There is damage to the optic nerve because of the backup of fluid in the eye.Glaucoma is a group of ocular disorders that are etiologically multifactorial and clinically associated with an optic neuropathy related to intraocular pressure (IOP).In the case of an infected eye, it can permanently damage vision, leading to blindness without treatment.An optic nerve injury results in the loss of visual (Pandey, B. K. et.al., 2021) perception, which is caused when light receptors in the brain are damaged.The number of Americans with Glaucoma is believed to be at least four million, and nearly half of those people are unaware that they are affected by glaucoma (Cupping, O. N., 2009).One of the best ways to prevent total blindness is to detect it early and prevent it.The detection of glaucoma diseases has been made possible by many techniques.For most of the diagnostic indexes like glaucoma, compiling optical disk (OD) information is critical.Based on the size of an image, researchers fix the radius of an iris based on the assumption that its center is closer to the center of an image.According to a region of interest (ROI), the optic disc is automatically detected from an eye image. Diagnosing glaucoma (Zhu, H. et. al., 2011) disease depends on the location of the optic disk.Before performing classification, the ROI is enhanced with image enhancement and features extracted.It is believed that there are a large number of edge detection filters available, each of which is capable of detecting particular types of edges (Shrivakshan, G. T., & Chandrasekar, C., 2012).The use of OCT for glaucoma screening has also been suggested in a few studies.An active contour approach was proposed by (Mishra, M., Nath, M. K., & Dandapat, S., 2011) to determine the glaucoma medicinal process from the color fundus images (Acharya, U. et.al., 2011).This method automatically extracts the ROI, reducing the time needed to detect the disease.Some CDR measures are used to extract the ROI in existing approaches (Krishnan, M. M. R., & Faust, O., 2013).We only utilize the features in this paper to detect the disease.The aim of this study is to explore the use of CNNs directly from unprocessed OCT volumes for the detection of glaucomatous eyes, i.e., bypassing the requirements of segmentation (Mookiah, M. R. K., 2012) to extract features (such as retinal layer thickness, rim volume, etc.. In contrast to classical machine learning methods with traditional segmentation-based features extracted from the same dataset of OCT scans, this CNN offers higher accuracy in classification (Muramatsu, C. et. al., 2010). Following are two approaches: feature-based and feature-agnostic approaches (Liu, Y. et. al., 2011). A segmentation-based approach uses an OCT volume as a feature-source; a feature-agnostic approach uses OCT volume as a feature (Manju, K. et.al., 2017) . In the early stages of glaucoma, patients usually do not complain of symptoms other than a small loss of vision in the peripheral field (Pachiyappan, A. et.al., 2012).

Figure 1.

Difference between normal and glaucoma affected eye