Design of a fingerprint template protection scheme using elliptical structures

Although biometric authentication is viewed as more prominent than password or token-based methodology in identity verification, biometric templates are vulnerable to attacks. This paper proposes a new approach for securing fingerprint templates using elliptical structures generated from the fingerprint minutiae. Authors generate a feature vector from the ellipse and will be projected onto a 3D-space to compute a binary string. The resultant binary string is transformed to frequency domain (DFT) and multiplied with a user specific random matrix to make it permanently non-invertible and secure. The results show the efficacy of the proposed method for protecting the fingerprints. (C) 2021 The Korean Institute of Communications and Information Sciences (KICS). Publishing services by Elsevier B.V.

Automated summary

The paper proposes a new method to secure fingerprint templates using elliptical structures generated from fingerprint minutiae. It involves generating a feature vector from the ellipse, projecting it onto a 3D-space, computing a binary string, transforming it to frequency domain (DFT), and multiplying with a random matrix to ensure security. The results demonstrate the effectiveness of the method in protecting fingerprints.