Multibiometrics architectures and privacy in a mobile environment

partners-icon-roma  Contact: Prof Patrizio Campisi

Objectives: Commercially available mobile devices have embedded a plethora of sensors able to acquire user based biometric information from face and fingerprint, to signature and iris. Moreover, the embedded inertial sensors in each mobile device allow data about gait and movement to the aforementioned list. Therefore, given the amount of available data, without adding any additional peripheral to the device, the opportunity to employ these signals for recognition purposes arises. Moreover, geo-localisation data and context meta-data represents soft biometrics that could be used in conjunction with the other data. In this framework this project deals with the design of a holistic system, fusing the information coming from the device for the purpose of obtaining a usable, secure, and privacy compliant multibiometrics system. The limited processing capabilities of a mobile device and higher processing time when compared to a server makes it necessary to design an optimised multibiometrics system architecture well suited to the mobile environment. Moreover, template protection techniques need to be properly designed and tested to operate within this framework. The three year project will be devoted to a) the analysis and implementation of different multibiometrics architectures b) the design of template protection algorithms for multibiometrics systems in a mobile scenario and c) usability analysis.

Expected Results: The project will have several expected outcomes. Specifically, the “optimal” architecture with respect to a mobile environment for fusing the data acquired through the device will be designed and tested. Its performance will be evaluated in terms of recognition capabilities and usability. Template protection mechanisms will be designed, implemented using a “privacy by design” approach.

Planned secondment(s): Two secondments will take place within this project. A three month visit to GENKEY will be carried out in order to gain insights about the industrial needs for the deployment of a multibiometric system in a mobile environment. Experiments using different mobile devices will be carried out to assess recognition performance and usability. The expected results of this secondment are: the design and implementation of multibiometric template protection architectures tailored to the limited processing of a mobile device. The product-oriented development competences of GENKEY will be instrumental to the aforementioned goal. The second secondment will be seven months with WUT where expertise in biometric algorithmic development and usability analysis will be leveraged. The expected results of this secondment are the usability assessment of the different designed multibiometric architectures based on expertise at WUT, particularly with respect to assessment and quality.

Start Date: October 2017