Detection of Diabetic Retinopathy using Multi-label Feature Extraction and Classification with Fully Homomorphic Encryption

Abstract

Diabetic Retinopathy (DR) is a circumstance that develops as a consequence of prolonged diabetes, directly impacting or implicating human vision. In its early stages, Diabetic Retinopathy (DR) typically presents no symptoms, and its progression can ultimately result in irreversible vision loss if not diagnosed and treated promptly. The computer supported the conclusion with the help of clinical pictures to help in provoke and exact treatment. Microaneurysms (MA) show the starting of DR, making it a significant calculate in diagnosing the malady. With the progression of Internet of Things (IoT), countless electronic gadgets are connected with the Internet. These associated electronic devices acquire and communicate data, and answer to any activities. In the medical system, hospitals can execute medical diagnosis (MD) with medical sensors, specially for inaccessible supporting MD. But, in this circumstance, patients’ privacy (PP) is of supreme importance, and privacy of medical data is decisive. Hence, the fundamental challenge ahead is the way to acknowledge distant assistant MD while safeguarding confidentiality of the clinical information and guaranteeing PP.