Iris Authentication Using Privatized Advanced Correlation Filter

This paper proposes a private biometrics formulation which is based on the concealment of random kernel and the iris images to synthesize a minimum average correlation energy (MACE) filter for iris authentication. Specifically, we multiply training images with the user-specific random kernel in frequency domain before biometric filter is created. The objective of the proposed method is to provide private biometrics realization in iris authentication in which biometric template can be reissued once it was compromised. Meanwhile, the proposed method is able to decrease the computational load, due to the filter size reduction. It also improves the authentication rate significantly compare to the advance correlation based approach [5][6] and comparable to the Daugmant’s Iris Code [1].