Security applications of distributed arithmetic coding

Distributed arithmetic coding (DAC) has recently been proposed for compression in the Slepian-Wolf setting. With respect to syndrome coding, DAC allows to easily adapt to non-stationary statistics of the signal to be coded, and works well for short and medium block lengths. In this paper we develop security applications of DAC in the field of biometric authentication. We show that DAC can be used for authentication by employing the codeword as secure hash, and using the outcome of DAC decoding for authentication. Moreover, we introduce a second powerful security feature, namely the randomization of the DAC intervals. This allows to protect the hash from attacks. We assess the authentication performance of the proposed scheme with respect to template matching and turbo codes.