FNeuroNet: privacy-focused medical image classification

Abstract

Artificial Neural Networks (ANNs) is increasingly used in healthcare to assist in early disease prognosis and decision-making; however, centralized training often faces challenges due to privacy regulations restricting the sharing sensitive patient data. In this study, FNeuroNet is proposed to facilitate decentralized training across multiple healthcare institutions by integrating federated learning with an advanced Convolutional Neural Network (CNN) architecture to mitigate this issue. FNeuroNet achieved an accuracy of 93.22 %, demonstrating robust performance while ensuring data security and compliance with relevant privacy standards.