Abstract: This paper defines some local minutiae structures around core region, and according to the matching results of these structures, the fiducial point is extracted accurately. In addition, a novel attribute, the number of ridges between minutiae and the fiducial point, is introduced and it is invariant to transition, rotation and nonlinear distortion. Minutiae pairs that are not matched obviously can be recognized by comparing the attribute of them in the matching stage, and the final decision is made with the method of a changeable sized boundary box. Experimental results demonstrate the efficiency of the matching algorithm is increased by 10%~20%, and a high matching precision is also achieved.

Key words: fingerprint matching, local structure, fiducial point, minutiae

摘要： 在图像中心区域为若干细节点分别构造局部结构，根据这些结构的匹配结果实现了基准点的精确定位。在指纹匹配阶段引入一个新的属性，即细节点到基准点之间的纹线切割数目，该属性对指纹的平移和旋转以及非线性形变具有不变性。指纹匹配时对该属性进行比较，排除明显不匹配的细节点，根据可变大小界限盒的方法对细节点进一步匹配。实验结果表明，该算法效率比原有算法提高了10%～20%，且实现了较高的匹配精度。

关键词: 指纹匹配, 局部结构, 基准点, 细节点

CLC Number: 

• TP18