Feature level fusion of fingerprint and finger vein biometrics

Abstract

The aim is to study the fusion at feature extraction level for fingerprint and finger vein biometrics. A novel dynamic weighting matching algorithm based on quality evaluation of interest features is proposed. First, fingerprint and finger vein images are preprocessed by filtering, enhancement, gray-scale normalization and etc. The effective feature point-sets are extracted from two model sources. To handle the problem of curse of dimension, neighborhood elimination and reservation of points belonging to specific regions are implemented, prior and after the feature point-sets fusion. Then, according to the results of features evaluation, dynamic weighting strategy is introduction for the fusion biometrics. Finally, the fused feature point-sets for the database and the query images are matched using point pattern matching and the proposed weight matching algorithm. Experimental results based on FVC2000 and self-constructed finger vein databases show that our scheme can improve verification performance and security significantly. © 2011 Springer-Verlag.