Lightweight, practical encrypted face recognition with GPU support

doi:10.1016/j.jisa.2026.104615

Lightweight, practical encrypted face recognition with GPU support

2026-09-15

0

PRE

AI

G

Gabrielle De Micheli *

S

Syed Mahbub Hafiz *

G

Geovandro C. C. F. Pereira *

E

Eduardo Lopes Cominetti *

DOI：10.1016/j.jisa.2026.104615

Original

Original request for help

Save

Abstract

En 中文

Face recognition typically operates in a client-server setting, where the client extracts a compact face embedding and the server performs similarity search over a template database. Since facial data is highly sensitive, this raises significant privacy concerns. Fully homomorphic encryption (FHE) addresses these concerns by enabling end-to-end encrypted similarity search. However, existing FHE-based protocols are computationally costly and, especially, impose high memory overhead due to large rotation-key sets and bandwidth-bound homomorphic operations. Building on prior work, HyDia (PoPETS 2025), we introduce algorithmic and system-level improvements targeting real-world deployment with resource-constrained (edge) clients. First, we propose BSGS-Diagonal, a fast and memory-efficient similarity computation algorithm that applies a Baby-Step/Giant-Step strategy with precomputed rotations reused across consecutive matrix–vector products. This yields a 91% reduction in rotation keys ( ∼ 14 GB less client memory) and cuts peak server-side CPU RAM usage from over 33 GB to 11 GB for databases up to 1M entries, with runtime improvements of up to 1.57 ×  for membership verification and 1.43 ×  for identification. Second, we introduce GPU-optimized similarity computation kernels, including an efficient homomorphic Chebyshev evaluator built upon FIDESlib (ISPASS 2025), a CKKS-level GPU library based on OpenFHE. Rather than offloading individual CKKS primitives, our integrated kernels fuse operations to avoid repeated CPU–GPU ciphertext movement and costly FIDESlib/OpenFHE data-structure conversions. Our HyDia and BSGS GPU results achieve up to 9 ×  and 21 ×  speedups on single GPU (and up to 287 ×  and 211 ×  using multi-GPUs), respectively, enabling sub-second encrypted face recognition for databases up to 2¹⁶ entries (or 2¹⁹ entries in a multi-GPU setting), while further reducing host memory usage.

Keywords:

Facial recognition

Privacy-Preserving biometrics

Fully homomorphic encryption

CKKS

GPU

Lightweight, practical encrypted face recognition with GPU support

Abstract

Journal

Organization

Cited Papers