電子情報通信学会ソサイエティ大会講演要旨
A-2-3
Biometric Identification System with Protected Templates under Noisy Enrollment
◎Vamoua YACHONGKA・Hideki YAGI(The Univ. of Electro-Communications)
We use the argument developed by Willems and Ignatenko to investigate the capacity region of the identification and secret-key rates of a biometric identification system with database consisting of protected templates. Compared to the previous study, we impose two more realistic conditions in our scenario; (i) we assume that the prior distribution of an identified user is unknown and (ii) there is a noisy discrete memoryless channel in the process of enrolling all individuals' bio-data sequence to the system database. The obtained result shows that the capacity region is similar to the one derived in previous research, and the unknown prior distribution does not decrease the capacity region.