RetinaCare: Diabetic Retinopathy Detection with VGG16-based Web Application

According to the International Diabetes Federation (IDF), According to the IDF, it is estimated that about one-third of people with diabetes have some form of diabetic retinopathy (DR). The prevalence of DR patients is projected to increase to 3.2 million by the year 2030. In high-income countries, there is an anticipated decrease in DR. Nevertheless, the focus persists on the vital significance of early identification and intervention in developing countries. The healthcare sector benefits significantly from early disease recognition. Advancements in artificial intelligence (AI) and deep learning (DL) in recent times have produced successful approaches in the detection and assessment of DR. Hence, the study introduces a web-based tool called RetinaCare, which integrates a CNN-based architecture for grading diabetic retinopathy (DR). To achieve multimodality, a CNN-based model is trained on top of VGG-16. The suggested framework exhibited an accuracy level of 96.89%, a specificity of 97.49%, and a sensitivity of 97.35%, showcasing its robust performance and high diagnostic capabilities. This innovative approach addresses the need for efficient and accessible solutions in resource-limited settings, contributing to advancing healthcare interventions for individuals at risk of diabetic retinopathy.