Revocable Crypto-Biometric Key Regeneration Using Face Biometrics, Fuzzy Commitment, and a Cohort Bit Selection Process

Abstract

Security is a main concern nowadays. Cryptography and biometrics are the main pillars of security. Using biometrics to obtain cryptographic keys offers distinct advantages over traditional methods. Classical systems rely on passwords or tokens assigned by administrators, which can be stolen or shared, making them insufficient for identity verification. In contrast, biometric-based keys provide a better solution for proving a user’s identity. This chapter proposes an approach to regenerate crypto-biometric keys with high entropy, ensuring high security using facial biometrics. The keys are regenerated using a fuzzy commitment scheme, utilizing BCH error-correcting codes, and have a high entropy of 528 bits. To achieve this, we use an intra-inter variance strategy for the process of bit selection from our facial deep binary embeddings. The system is evaluated on the MOBIO dataset and gives 0% FAR and less than 1% FRR. The proposed crypto-biometric keys are resistant to quantum computing algorithms, provide non-repudiation, and are revocable and convenient with low false rejection rates.