基于细节点聚类的多参考中心指纹匹配算法

Minutia Clustering-Based Fingerprint Matching Method Using Several Reference Points

指纹匹配算法的好坏直接影响识别系统的精度。提出了一种新的基于细节点聚类的多参考中心指纹匹配算法,在两枚指纹对齐阶段,不仅考虑了指纹的全局特性而且根据不同的细节点类自适应地构造不同的局部结构,有效地利用了一些孤立但信息量较大的细节点,提高重叠区域内细节点较少且分散的情况下对齐的准确性。在匹配阶段,多参考中心的使用和相似元分析的结合能在一定程度上克服指纹非线性形变的影响,降低了匹配算法的拒识率。实验结果表明该方法提高了匹配的性能。

Fingerprint matching algorithm is the main factor to affect the precision of a fingerprint verification identification system. A new fingerprint matching method using several reference points based on minutia clustering is proposed in this paper. At the alignment stage, the matching algorithm combines the global structures of minutia and the local structures which are constructed in a variety of type according the number of minutia in a cluster adaptively. The method can take advantage of the most of some isolated minutiae which includes some discriminative information for aligning fingerprints what a small number of minutia are scattered in a large area, and it definitely improve the accuracy of the alignment. At the matching stage, using several reference points and in combination with similar element analysis alleviate the effect of nonlinearly deformation and decrease the false reject rate. Experimental results show that the proposed algorithm improves the performance of the fingerprint matching.