EuroGP'06 Proceedings of the 2006 international conference on Applications of Evolutionary Computing
| Hi-index | 0.00 |
Fingerprint matching is still a challenging problem forreliable person authentication because of the complexdistortions involved in two impressions of the same finger.In this paper, we propose a fingerprint matchingapproach based on Genetic Algorithms (GA), which findsthe optimal global transformation between two differentfingerprints. In order to deal with low quality fingerprintimages, which introduce significant occlusion and clutterof minutiae features, we design the fitness function basedon the local properties of each triplet of minutiae. Theexperimental results on National Institute of Standardsand Technology fingerprint database, NIST-4, not onlyshow that the proposed approach can achieve goodperformance even when a large portion of fingerprints inthe database are of poor quality, but also show that theproposed approach is better than another approach,which is based on mean-squared error estimation.