[en] Sometimes fingerprint-like features are found in a material. The exciting discovery poses new challenges on how to use the features to build an object authentication protocol that could tell customers and retailers equipped with a mobile device whether a good is authentic or fake. We are exactly in this situation with Cholesteric Spherical Reflectors (CSRs), tiny spheres of liquid crystals with which we can tag or coat objects. They are being proposed as a potential game-changer material in anti-counterfeiting due to their unique optical properties. In addition to the problem of processing images and extracting the minutiæ embedded in a CSR, one major challenge is designing cryptographically secure authentication protocols. The authentication procedure has to handle unstable input data; it has to measure the distance between some reference data stored at enrollment and noisy input provided at authentication. We propose a cryptographic authentication protocol that solves the problem, and that is secure against semi-honest and malicious adversaries. We prove that our design ensures data privacy even if enrolled data are leaked and even if servers and provers are actively curious. We implement and benchmark the protocol in Python using the Microsoft SEAL library through its Python wrapper PySEAL.
[University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > IRiSC >]
