GFT-GCN: Privacy-Preserving 3D Face Mesh Recognition with Spectral Diffusion
Published in WACV 2026, 2026
3D face recognition provides strong security but requires protection of stored biometric templates. We propose GFT-GCN, a privacy-preserving framework that combines spectral graph learning and diffusion-based template protection to generate secure, irreversible templates. Experiments on BU-3DFE and FaceScape show high accuracy and strong resistance to reconstruction attacks, achieving a good balance between privacy and performance.