FHE-Based Secure Image Processing Framework with Integrated Key Management System
Rafik Hamza · 2024
Image processing on encrypted data presents significant challenges, requiring advanced computations and robust access control mechanisms to ensure data security. Fully Homomorphic Encryption (FHE) emerges as a promising technology in this area and applicable to various real-world scenarios. However, without a comprehensive security framework governing access to encrypted images, practical adoption of FHE remains limited. In this paper, we propose a secure framework for encrypted image processing that seamlessly integrates FHE technology with a Key Management System (KMS). Our framework enables users to perform diverse image analysis operations on encrypted datasets, maintaining different layers of authentication without sharing their secret keys. We validate our framework by a simple implementation using the TFHE library and the HashiCorp Vault KMS, demonstrating both its speed and effectiveness for secure image processing tasks. Additionally, we conduct a thorough security analysis, assessing its resilience against various attacks. Our findings underscore the framework’s efficacy in safeguarding sensitive and private images with a robust security framework.