Logo des Repositoriums

Towards improving the NIST fingerprint image quality (NFIQ) algorithm


The NIST Fingerprint Image Quality (NFIQ) algorithm has become a standard method to assess fingerprint image quality. However, in many applications a more accurate and reliable assessment is desirable. In this publication, we report on our efforts to optimize the NFIQ algorithm by a re-training of the underlying neural network based on a large fingerprint image database. Although we only achieved a marginal improvement, our work has revealed several areas for potential optimization.


Merkle, Johannes; Schwaiger, Michael; Breitenstein, Marco; Bausinger, Oliver; Elwart, Kristina; Nuppeney, Markus (2010): Towards improving the NIST fingerprint image quality (NFIQ) algorithm. BIOSIG 2010: Biometrics and Electronic Signatures. Proceedings of the Special Interest Group on Biometrics and Electronic Signatures. Bonn: Gesellschaft für Informatik e.V.. PISSN: 1617-5468. ISBN: 978-3-88579-258-1. pp. 29-44. Regular Research Papers. Darmstadt. 09.-10. September 2010