Gaussian Differential Privacy Federated Learning to Identify Ocular Diseases
摘要
Conditions affecting the visual system and eyes are known as ocular diseases and it may lead to health problems, if untreated. Thus, to prevent from the health problem, it is crucial to attend for treatment. The ocular disease data is utilized in this research work for prediction of visual problem using federated learning. Protecting sensitive and personal information, particularly eye images, is crucial. Further, to prevent misuse and protect an individual’s privacy and security, this research work applies federated learning model. Therefore, the proposed framework guarantees that federated learning model is safely and discreetly trained on data using Gaussian noise technique, which would aid in the identification of ocular diseases and improve the privacy between global and local communication.