Throughout the last decade, several major security breaches have been the result of weak passwords chosen for the benefit of memorability. Biometric data on the other hand does not suffer from this requirement to be memorable, but its immutability requires additional effort in secure storage and processing compared to passwords.
Following improvements in the area of cryptography, research has been targeted towards these requirements but was always restricted by the available tools of its time in securing biometric access control. This work aims to close gaps. We provide a semi-honestly secure solution in the case of biometric templates w.r.t. a dissimilarity measure. In particular, our solution never decrypts sensible data during the entire computation process. Relying on fully homomorphic encryption, the newly developed method of programmable bootstrapping is crucial to achieve a feasible run time. The security proof given by us can be translated to all biometric access control systems meeting certain conditions.
Doing so, we believe to have developed a black box solution for securing biometric access control that tremendously reduces the probability of adversaries to acquire biometric data. Similarly, we assume the constructed framework for the security proof to be an enabler regarding future work in the development of more efficient systems. We believe that this is a huge step in closing the security gap between biometric data and strong passwords making the prior a better choice for usable security for access control in the near future.
The thesis has won a BIG thesis award donated by the Institute of Computer Science Bonn and the Bonner Informatik Gesellschaft.