Biometrics Security & Privacy

Biometrics refers to the automatic recognition of individuals based on their physiological and/or behavioral characteristics.

The Biometrics Security and Privacy (BSP) group at Idiap focuses on three main areas of research:

1. Biometric recognition: We investigate and develop new biometrics based recognition algorithms, notably for face, voice, and vein biometric modalities.
   
2. Presentation attack detection: We look for new and better ways of detecting presentation attacks on face, voice, and vein biometric recognition systems.
   
3. Biometric template protection: We research effective methods of preserving both the security of biometric recognition systems and the privacy of their users by protecting the biometric models ("templates") that are employed by the system for recognition purposes.
   

The BSP group prioritizes reproducibility in research. This is important for ensuring that our work can be both verified and built upon by the wider research community. To enable reproducibility, we mainly make use of our Python-based signal-processing and machine-learning toolbox, Bob, which we make freely available for academic purposes. The group also develops and maintains the BEAT platform -- a MLaaS platform compliant with Swiss and European data-security norms.

The group participates in several large-scale biometrics projects at Swiss (SNSF), European (H2020) or world-wide levels (eg. IARPA/DARPA) but also conducts projects directly with companies.

The BSP group provides also expertise to the Swiss Center for Biometrics Research and Testing, which, among other things, carries out tests and evaluations on commercial products related to biometrics.

More News

More Jobs