Fiche de IBARRONDO LUIS Alberto

Titre : Privacy-preserving Biometric Systems with Advanced Cryptographic Techniques
Date : 27 mars 2023
Heure: 10:00
Résumé : Many research works have been proposed to address privacy protection of both the biometric data involved and the models behind it. Most privacy preserving neural networks either use secure multi-party computation (MPC), homomorphic encryption (HE) or a combination of both. Employing these two techniques, as well as Functional Encryption (FE), this thesis tackles the design and implementation of practical privacy-preserving biometric systems. The thesis aims to find the proper balance between privacy, accuracy and performance, and to provide solutions for multiple scenarios that are both secure and efficient, dealing with all the aspects that incur in accuracy losses. Regarding the threat model, we go beyond protections against semi-honest adversaries (preserving privacy) by also addressing malicious adversaries (preserving correctness). Last but not least, we address the leakage of biometric data when revealing the output, a privacy concern often overlooked in the literature.
Lieu : EURECOM Campus SophiaTech, 450 Route des Chappes, 06410 Biot, France