Comparative analysis of homomorphic encryption schemes for encrypted image processing in OpenStack using TenSEAL

Saker, Hekmat and Eyvazov, Farid and Emad Ali, Tariq and Biró, Csaba and Dhulfiqar, Zoltán Alwahab (2025) Comparative analysis of homomorphic encryption schemes for encrypted image processing in OpenStack using TenSEAL. Annales Mathematicae et Informaticae, 62. pp. 123-142. ISSN 1787-5021 (Print) 1787-6117 (Online)

Preview

Text AMI_62_from123to142.pdf - Published Version Download (828kB) | Preview

Abstract

This thesis presents a detailed comparative analysis of several prominent homomorphic encryption (HE) schemes-BGV, CKKS, and TFHE for secure encrypted image processing within an OpenStack cloud environ ment utilizing the TenSEAL library. With cloud computing becoming in creasingly prevalent, ensuring data privacy during remote data processing has emerged as a critical challenge. Image processing, particularly bright ness adjustment, is chosen as a representative task to explore the practicality and computational efficiency of HE schemes. Experiments conducted on a simulated cloud setup revealed significant differences in computational perfor mance, memory utilization, and practical feasibility. Specifically, the CKKS scheme exhibited superior efficiency for approximate arithmetic operations, making it particularly well-suited for real-number computations in image pro cessing. These findings highlight the strengths and limitations of current HE schemes, providing valuable insights into their practical deployment in cloud environments, with implications for improved data security and privacy.

Actions (login required)

Edit Item