Performance and security‐enhanced fuzzy vault scheme based on ridge features for distorted fingerprints

Abstract

The fuzzy vault scheme is a well-known technique to protect the fingerprint templates and its secret key simultaneously in biometric systems. The performance of the fuzzy vault scheme is affected if collected fingerprint images are distorted. It is proved that ridge features of a fingerprint image have the advantage of being invariant to any geometric transformations. Therefore, the authors proposed a novel fingerprint fuzzy vault scheme based on ridge features information with the goal of improving its performance for distorted fingerprint images. They also proposed a novel noise generation technique to improve the security of the proposed fuzzy vault scheme. Experiments are carried on FVC2002 and FVC2004 databases, which include distorted fingerprint images. When the key size is 160 bit, the genuine accept rate is 89 and 75% with 0% false accept rate on FVC2002-DB2A and FVC2004-DB3A, respectively. Experimental results show that the proposed scheme achieves better performance than the existing methods. The security level of the proposed system is evaluated based on the min-entropy analysis and the degree of freedom of generated noise points. The security of their proposed fuzzy vault scheme is ?42 bits. This proves that the proposed scheme maintains the security when dealing with distorted fingerprint images.