Minutiae-Based Template Synthesis and Matching Using Hierarchical Delaunay Triangulations

Fingerprint-based authentication is a key biometric technology with a wide range of potential applications both in industry and government. However, the presence of intrinsically low quality fingerprints and various distortions introduced during the acquisition process pose challenges in the development of robust and reliable feature extraction and matching algorithms. Our focus in this study is on improving minutiae-based fingerprint matching by effectively combining minutiae information from multiple impressions of the same finger. Specifically, we present a new minutiae template-merging approach based on hierarchical Delaunay triangulations. The key idea is synthesizing a super-template from multiple enrollment templates to increase coverage area, restore missing features, and alleviate spurious minutiae. Each minutia in the super-template is assigned a weight representing its frequency of occurrence, which serves as a minutiae quality measure. During the merging stage, we employ a hierarchical, weight-based, scheme to search for a valid alignment between a given template and the super-template. The same algorithm, with minor modifications, can be used to compare a query template with a super-template. We have performed extensive experiments and comparisons with competing approaches to demonstrate the proposed approach using a challenging public database (FVC2000 Dbl).