Secure Fingerprint Recognition Based on Frobenius Norm

Aiming at the variance of fingerprint feature in multi-capturing and the security of fingerprint template, research on secure fingerprint recognition system based on Frobenius norm is given in this paper. Frobenius norm can show the main information in region of interest (ROI). The random mapping results can improve effectively the distinguishment of different users. Compared with the original fingerprint feature being stored in the database, the mapping results can avoid attacking the original fingerprint feature. For the given fingerprint database, the optimal threshold and the optimal size of random mapping matrix can be got. Experiments in FVC2002 show that the proposed recognition system is feasible and effective. Moreover, a satisfactory EER (equal error rate) 0.21% is obtained.