A new fingerprint matching algorithm based on minimum cost function

Abstract:

We develop new minutia-based fingerprint algorithms minimizing a cost function of distances between matching pairs. First, using the minutia type or minutia quality, we choose a reference set of points form each set. next, we create the set of combinations of pairs to perform the best alignment and finally the matching by distances is computed. We tested our algorithm using the DB2A FVC2004 database extracting the minutia information by the mindtct program given by NBIS and we compare with the bozorth3 algorithm performace. © 2009 Springer-Verlag.

Año de publicación:

2009

Keywords:

Fuente:

scopus