BFV-Based Homomorphic Encryption for Privacy-Preserving CNN Models

Medical data is frequently quite sensitive in terms of privacy and security. Federated learning has been used to increase the security medical data, which a sort machine technique. The training disseminated across numerous machines federated learning, process collaborative. There are attacks on deep (DL) models that attackers can use obtain information. As result, DL model should be safeguarded from adversarial attacks, particularly applications. Homomorphic encryption-based collaborator one answers this challenge. Using homomorphic encryption, research presents privacy-preserving system for data. proposed technique employs secure multi-party computation protocol safeguard adversaries. approach tested performance using real-world dataset paper.