Securefedlung: Privacy-Preserving Multi Disease Diagnosis Using Federated CNN and ECC

Abstract

Medical image analysis plays a crucial role in early disease detection and diagnosis in modern healthcare systems. Deep learning techniques, particularly Convolutional Neural Networks (CNN), have significantly improved the accuracy of medical image classification for detecting diseases such as lung cancer, pneumonia, and tuberculosis. However, traditional centralized learning approaches require sharing sensitive patient data, raising serious privacy and security concerns. This project proposes SecureFedLung, a privacy-preserving multi-disease diagnosis framework that integrates Federated Learning (FL), Convolutional Neural Networks (CNN), and Elliptic Curve Cryptography (ECC). In this system, multiple hospitals collaboratively train a global model without sharing raw patient data. Each hospital trains a local CNN model and shares only encrypted model parameters using ECC. The central server aggregates these encrypted updates using the Federated Averaging algorithm to generate a global model. The system ensures secure communication, protects sensitive medical data, and achieves high diagnostic accuracy. Experimental results demonstrate that SecureFedLung provides an efficient, secure, and scalable solution for collaborative medical AI systems.