Advances in the Application of Convolutional Neural Networks for Glaucoma Diagnosis

Abstract

Glaucoma is one of the main diseases that cause blindness. It is a retinal eye condition that is harmful, takes a long time to manifest, and affects an increasing number of people. An effective method for the early detection of glaucoma is the use of a computer-aided diagnostic system, which allows medical professionals to assess the condition with results that often match or exceed those of human judgement. As a consequence of the deep learning revolution and its resounding win over conventional machine learning (ML) techniques, several DL methodologies have been presented. In this article, we review the advantages and disadvantages of using retinal images to assess the progression and categorization of glaucoma using convolutional neural networks. The available retinal image datasets and performance assessment measures are reviewed. This review aims to draw attention to gaps and directions for future research.