Description
Date depot: 23 juin 2020
Titre: Biometric recognition algorithms with secure multiparty computation
Directrice de thèse:
Melek ONEN (Eurecom)
Directeur de thèse:
Hervé CHABANNE (LTCI (EDMH))
Domaine scientifique: Sciences et technologies de l'information et de la communication
Thématique CNRS : Non defini
Resumé:
The goal of the PhD is to investigate recent advances in privacy enhancing technologies and further design and evaluate privacy preserving variants of biometric classification algorithms that will on the one hand protect the confidentiality of the data and on the other hand enable authorized third parties to perform operations over the encrypted data. To this end, the privacy and security challenges associated with biometric classification based on convolutional neural networks (CNNs) will be studied. Recent advances on cryptographic primitives and in particular secure multi-party computation (SMC) will further be investigated. The ultimate goal is to integrate SMC with biometric recognition algorithms taking into account the complexity of state of the art CNNs. These CNNs count millions of parameters [6] and how to securely implement them in this distributed environment remains challenging.
Doctorant.e: Ibarrondo Luis Alberto