An enhanced fuzzy vault to secure the fingerprint templates

摘要

Fingerprint-based biometric systems have significant advantages over the conventional authentication systems, which are based on passwords and tokens. However, these systems are needed to combat the increasing magnitude of identity theft of users enrolled in a fingerprint-based biometric system because the fingerprint information of a user cannot be changed if it is compromised. Moreover, it has been demonstrated in the literature that a fingerprint image can be reconstructed if the information of minutiae points is available. In this paper, a fuzzy vault based technique is proposed to prevent identity theft and secure the fingerprint information (essentially, minutiae points) stored in the database. We propose a novel technique to filter the genuine vault points from a combination of genuine and chaff points used in the fuzzy vault technique. Since minutiae points are used to construct the vault, it is a challenging task to align probe and gallery images during verification. In order to do that, a Principal Component Analysis (PCA) based alignment technique is also proposed to align the gallery and probe templates. The proposed technique is evaluated on three different Fingerprint Verification Competition (FVC) databases that come under the FVC2002 and FVC2004. Subsequently, the obtained results are compared with that of the recent existing techniques in the literature and are found to be superior in terms of the Genuine Acceptance Rate (GAR), False Acceptance Rate (FAR), and Equal Error Rate (EER).