Abstract
Diabetic Retinopathy (DR) and Age-related Macular Degeneration (AMD) can prevent or reduce the severity of the disease if the diseases are found early. This can lead to a treatment to reduce loss, reduce medical expenses, and increase the quality of life for the elderly. Diagnosis of DR and AMD must be examined by an ophthalmologist by retinal examination. However, ophthalmologists in Thailand are still inadequate for their needs, especially in remote areas. With such restrictions, patients do not receive proper treatment on time and that can cause vision loss. Therefore, our research team developed DeepEye using deep learning, a technique in artificial intelligence that is very accurate. By developing it to be easy to use in the form of application on smartphone that can connect to any portable fundus cameras, can be screened quickly, and provide very high accuracy equivalent to an ophthalmologist. Specifically, it can be used to screen DR and classified into three types of retinal images that are normal, DR, and other diseases. If detecting a diabetic retinopathy, DeepEye can also accurately identify the severity of DR into 4 levels, i.e., mild, moderate non-proliferative DR (moderate NPDR), severe (severe non-proliferative DR: severe NPDR), and severe with new blood vessels (proliferative diabetic retinopathy: PDR) which is a very dangerous stage of DR. Moreover, DeepEye can be used to screen AMD, which can be classified into dry age-related macular degeneration (dry AMD) and wet age-related macular degeneration (wet AMD).
