“Bubble Tag”-based system for object authentication

Biometric systems are omnipresent nowadays in fields that require user authentication (e.g.: access control, banking operations), due to the main attributes of the biometric characteristics: uniqueness, permanence, collectability. In the products world, a solution that could achieve the same performance as a biometric system could be represented by a “Bubble Tag”-based system. In this article we propose solutions for signature extraction and for “1 to many” authentication protocol applicable to the Bubble Tag. We briefly present a signature extraction method that is invariant under perspective. For the “1 to many” authentication protocol we recommend the use of an LSH (locality sensitive hashing) approach. Tests carried out on randomly computer-generated images gave promising results and indicated leads to be followed for real images.