We are proud to announce that we presented at the DIGITAL 2025 conference in Venice, Italy, where we introduced our latest research on the challenges of protecting facial biometric templates.
The conference organizers recognized our talk and the related paper with a Best Paper Award. The recognition encourages us to keep innovating in the protection of face recognition.
What is the paper about?
The paper looks at how to store and process securely the so-called face embeddings used by face recognition systems, that is, the machine-readable features computed from faces.
These biometric templates play a key role in modern security and commercial systems. At the same time, they carry serious data protection risks, especially under strict regulations such as the GDPR or the AI Act. The paper compares two main technologies and, on that basis, proposes directions for further research. As part of the research, we also present a practical case study based on the Labeled Faces in the Wild face recognition dataset. We also suggest a few new optimization options for homomorphic encryption approaches.
The goal is to develop biometric systems that are not only effective but also sustainable from a data protection point of view.
Why does this matter to us?
Because we have built these research results (and much more) into our face recognition service, to make it safer for everyone involved. If you would like to use face recognition in your own service, take a look at our VeriDome service and get in touch.
Related links: