A statistic feature-based scheme for the template protection of 2D face recognition

The security of biometric template is important for a biometric system's privacy and security. In a traditional biometric system, the template is saved directly in a reference database. This results in a series of security and privacy risks, which limits the development of biometric recognition technique. This paper provides a template protection scheme that is suit for statistic feature-based algorithm, which has been widely used in face recognition. The core idea of the scheme is the match between template and tested feature by using Hash function. The scheme includes the following four parts, which are feature extraction, primary components selection, quantization and error correction coding, to eliminate the contradiction between the fuzziness of the biometric information and the sensitiveness of the hash function, and the contradiction between the short feature vector obtained by statistic feature extraction algorithm and the system security. The proposed scheme has been applied to 2D face recognition, the analysis demonstrated the security of the scheme and the simulation results showed the feasibility of the scheme.