Face template protection using chaotic encryption

With the wide spread of biometric system and growing concerns about personal privacy, the security of biometric template itself is more important for a biometric system. Chaotic systems have significant properties, such as pseudorandom property, ergodicity, sensitivity to initial conditions and system parameters etc. These make it suitable for image encryption and there are many applications adopted chaotic encryption in this area. In this paper, we propose a template protection algorithm for face recognition adopted chaotic cryptographic technology. The original template is created by Fisherface, then scrambled by the chaotic sequences generate based on logistic map in order to protect the privacy of original template. Experimental results on ORL database show that, the LDA features disturbed utilize Logistic map can ensure the security of the original template and the performance of the system is not significant degraded.