A Robust and Efficient Minutia-Based Fingerprint Matching Algorithm
Wen Wen,Zhi Qi,Zhi Li,Junhao Zhang,Yu Gong,Peng Cao
DOI: https://doi.org/10.1109/ACPR.2013.43
IF: 8
2013-01-01
Pattern Recognition
Abstract:In this paper, we propose a novel robust and efficient minutia-based fingerprint matching algorithm. There are two key contributions. First, we apply a set of global level minutia dependent features, i.e., the qualities that measure the reliabilities of the extracted minutiae and the area of overlapping regions between the query and template images of fingerprints. The implementation of these easy-to-get minutia dependent features presents coherence to the well-accepted fingerprint template standards. Besides, the reasonable combination of them results in the robustness to poor quality fingerprint images. Second, we implement a hierarchical recognition strategy, which applies a procedure of global matching that refines the local matching decision towards a genuine result over the entire images. Other than the much improved accuracy, our algorithm also promotes the efficiency, because compared with other state-of-the-art matching approaches, it does not make use of any time-consuming operations or any complex feature structures. The experimental results demonstrate the proposed method exhibits an excellent accuracy that exceeds the performances of well-known minutia based matchers. Meanwhile, the proposed algorithm presents potentials to serve a real-time fingerprint recognition system.