Detection of Diabetic Retinopathy With Mobile Application Using Deep Learning

Introduction

Technology made our life easier by facilitating our work in most areas by bringing new innovations every day. Although it has been suggested that the advancement of every technological development brings along various damages, technology makes daily life easier for people and finalizes problems that cannot be solved or even partially solved by certain and existing methods. These advantages have increased the place and importance of technology for humanity and have brought a new dimension to the functioning of our age and human life.

Recently, in addition to the development of computer and communication technologies, the technology area called artificial intelligence is also receiving a lot of attention. Artificial intelligence is known as offering or exploring science field by using algorithmic solutions to problems based on the behavior of humans and living beings (Russell, 2016).

Increasing technological developments with acceleration have made it inevitable to be included in our life and seen and used math-based systems such as artificial intelligence, image processing, machine learning, and deep learning. Undoubtedly, the medical world comes first among them (Foster, 2014). Artificial intelligence and its fields provide the opportunity to obtain an early diagnosis and give faster results for the treatment of disease (Karaboğa, 2011). Studies on this area are increasing and developing day by day.

The fact that artificial intelligence in especially human health is used for the diagnosis of disease and produces good results and also gives good signs for the research of diabetic retinopathy and causes the studies on this subject to become widespread. Therefore, bringing to the literature new studies is very important.

There are almost 415 million diabetic patients in the world. It was reported that approximately 285 million of them have diabetic retinopathy, and the 40 to 45 million diabetic patients who have diabetic retinopathy are seriously threatening to see (Christian Nordqvist, 2017). Diabetic retinopathy is one of the leading causes of blindness in the world (Christian Nordqvist, 2017). Therefore, ophthalmologists suggested that diabetic patients need a more frequent eye test. With the early diagnosis of diabetic retinopathy; it is possible to prevent blindness as a result of diabetes and to improve the patient's vision. According to studies, it is predicted that patients with diabetic retinopathy will reach 370 million by 2030. Although these numbers are very serious and undeniable, to parallel the studies conducted with the development of technology aim to raise the awareness of the potentially risky patients, as well as the patients who have been diagnosed, and to reduce the number of patients.

Symptoms of diabetic retinopathy may be evident by the patient only at an advanced stage, but an ophthalmologist can detect the symptoms before reaching this stage. It is very important that diabetic patients have eye test at least twice in a year. In this way, the chances of early diagnosis and treatment of the diabetic retinopathy increase. Most commonly used methods detecting the diabetic retinopathy are expanded eye examination, fluorescein angiography (FA), and optical coherence tomography (OCT) (Clairhurts Eye Care, 2019).

Figure 1.

Ophthalmoscope

(Clairhurts Eye Care, 2019)

Figure 2.

Optical coherence tomography (OCT)

(Eye Care, 2019)

Figure 3.

Image of healthy retina photo

(IEEE DataPort, 2019)

Figure 4.

Image of retinal disease

(IEEE DataPort, 2019)

Figure 5.

Image of retinal disease

(IEEE DataPort, 2019)