Biometric Identification Systems with Both Chosen and Generated Secrecy

We investigate the fundamental limits of a combined usage of chosen- and generated-secret keys for biometric identification systems with exponentially many users. The system consists of two phases: enrollment and identification phases. In the enrollment phase, for each user, the encoder generates a secret key and a helper data by using another secret key, chosen independently, and the bio-data sequence. In the identification phase, observing biometric data sequence, the decoder estimates index, chosen- and generated-secret keys of the identified user. In this study, the capacity region of identification, chosen- and generated-secrecy, template, and privacy-leakage rates of the system is characterized. The characterization shows that there is a trade-off among identification, chosen- and generated-secrecy rates, and greater values of identification and chosen-secrecy rates require larger storage space, but the generated-secrecy rate does not contribute to such increase. Also, this result includes various known results as special cases.