ESR 3: Template protection in biometric-based mobile scenarios

Objectives: Privacy is becoming more and more a key issue that needs to be addressed for the acceptance of biometric systems by the general public. The design of privacy enhancing technologies (PET) tries to answer this need by transforming the biometric reference information, namely the template, in identifiers that ideally leak little or no information about the underlying biometric that was captured, thus minimising privacy issues. Furthermore, comparison is carried out directly in the transformed domain of the identifier space, thus allowing the original template to be discarded. With the embedding of biometric sensors in mobile devices and their use in everyday applications, the need to use template protection techniques properly designed for mobile applications, arises. This problem has not been tackled in literature to date. This three-year project will deal with the design of template protection algorithms specifically tailored to the mobile scenario. Specifically our activity will be devoted to a) the design of biometric cryptosystems, that is systems that combine signal processing tools and cryptographic primitives, b) transformation based template protection approaches, that is systems that perform non-invertible transformations on the original template and c) defining effective metrics for security assessment of a protection scheme with respect to compliance with the required properties such as non-invertibility.

Expected Results: The project will have several expected outcomes. Specifically, a better understanding how the mobile scenario changes the paradigm for template protection will be gained. Appropriate template protection techniques will be implemented in the mobile scenario and they will be evaluated through proper metrics to be proposed in this project.

Planned secondment(s): Two secondments will take place within this project. The first is a three month visit to GENKEY where the development of biometric template protection algorithms will be explored, and implemented for a mobile scenario within an industrial context. The expected results of this secondment are: the design and implementation of template protection algorithms relying on the well consolidated experience of GENKEY on template protection specifically on the algorithm industrial implementation. The second seven-month secondment will be with OVGU wherein the ESR would further develop and test the metrics for assessing the template protection method. The expected results of this secondment are: the definition of metrics properly designed for the mobile environment and their test against the irreversibility and unlinkability properties leveraging on OVGU expertise on privacy management.

Start Date: November 2017