Abstract

ESTEVEZ BRESO, David  and  MUNOZ BRISENO, Alfredo. Minutiae template fusion and its impact on fingerprint matching. Rev cuba cienc informat [online]. 2018, vol.12, n.4, pp. 103-114. ISSN 2227-1899.

In this work, a new minutiae template fusion algorithm, obtained from fingerprints is proposed. This algorithm finds initial correspondences between minutiae, using an accurate state of the art algorithm (MCC). Correspondences that give as result the geometric transformation that best aligns the minutia templates are chosen, using a novel process. For this, a new heuristic that refines the result by finding the “feasible” transformation that involve more correspondences, is introduced. In this way, a new algorithm that is based on measuring the provoked effect in terms of distances between correspondent minutiae when a new correspondence is added to the transformation computation, is proposed. The approach is very useful when there are many occluded impressions of the same finger, or when the captured impressions have disjoint areas. The positive impact of the proposed algorithm in the accuracy of the fingerprint identification is tested by using two different matching algorithms. The results were obtained database FVC 2004 DB1_A and in two variants of this database.

Keywords : minutiae based template fusion; point set registration.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )