A Multiple Feature Fusion System for Fingerprint Recognition

Abstract

A number of different features, besides minutiae, have been used for fingerprint matching. Previous studies have shown that the performance of a fingerprint recognition system can be improved by combining these features with minutiae through a fusion strategy. However, most of these studies extract only a single type of feature for fingerprint recognition, and then fuse it with conventional minutiae-based method to improve performance. In this paper, in order to select the best fusion of fingerprint features, a comparative study of combining multiple features is firstly carried out on several fingerprint databases. All the comparing results show that, generally the more features being used, the better the performance is. However, beyond four features, the performance improvement is negligible. Based on these observations, a multi-feature based fingerprint recognition system using the best combination of the four features is proposed. The experimental results comparing to the-state-of-art algorithm show the effectiveness of the proposed system.