高级检索

    基于Delaunay三角化的指纹匹配方法

    Fingerprint Matching Based on Delaunay Triangulation

    • 摘要: 将计算几何的三角划分方法引入指纹匹配,研究了一种基于DT(Delaunay triangulation)网的指纹匹配方法.通过对细节点的拓扑结构进行DT划分,把空间上位置相近的细节点按照一定规则相连,得到三角形网格.然后基于该网格寻找若干参考点对,并根据获得的参考点对将两幅指纹图像进行姿势调整.最后使用获得的参考点对实现基于点模式的指纹匹配.算法在第1届中国生物特征识别竞赛指纹组的测试结果证明了有效性.

       

      Abstract: In this paper, a method for fingerprint matching is proposed based on Delaunay triangulation in computational geometry. First, minutiae taken from the template and the query fingerprint images are triangulated using the Delaunay's rule. Then reference minutiae pairs are obtained by searching two DT nets. Finally the query fingerprint image is adjusted according to the template fingerprint image with parameters computed from reference minutiae pairs and match score is calculated using a simple match algorithm. The experiments conducted on BVC2004 confirm the effectiveness of the proposed algorithm.

       

    /

    返回文章
    返回