A Review on the Detection of Diabetic Retinopathy through the Use of Deep Learning

Abstract

One of the most common and complicated forms of diabetes, diabetic retinopathy (DR) is now recognised as the major cause of blindness on a global scale. It is also one of the most commonly observed forms of the disease. The development of DR is associated with a degeneration of the blood vessels in the retina. Over the course of the past few years, numerous approaches to DR detection have been put forward as a result of the growing recognition that an early diagnosis is critical to effective treatment. However, research done in recent times has revealed a fact that deep learning-based CNN structures and Transfer Learning-based Medent’s are widely employed in DR detection because of their superior performance in the medical sector. This is a reality that has been brought to light as a result of recent developments. This article presents a study on automated ways that are used to diagnose diabetic retinopathy utilising image processing and disease classification techniques as a result of such improvements in Deep Learning methodology. These advancements have led to the writing of this article.