基于指纹结构特征信息匹配的算法

Fingerprint Matching Algorithm based on Structural Features Information

为了克服指纹识别中常见的问题,提出了一种基于指纹结构特征信息匹配的算法。该算法利用改进的Bresenham算法求得指纹分叉点间连线所穿越的脊线个数和分叉点结构特征信息,得到模板指纹和待识指纹的结构特征信息矢量数组;运用改进的二分图的完美匹配算法,得到矢量数组的匹配度。对该匹配度进行评估,如果高于某一个阈值,则认为指纹匹配成功;否则,则认为不是同一指纹。该算法计算速度快,具有较高的识别率,且能较好地正确识别扭曲、偏移、信息残缺等质量较差的指纹图像,在实际应用中取得了较好的效果。

In order to overcome the common problems in fingerprint recognition, an algorithm based on the structural information is presented in this paper. The .sums of through ridges and structural feature information of bifurcates are extracted by using the developed Bresenham algorithm, and the vector arrays of structural information of template fingerprint and recognized fingerprint are computed; the matching degree of vectors is gained in the way of developed perfect matching algorithm of bipartite graph. The matching is success if the degree is bigger them a certain threshold, or the fingerprints are not from the same person. The algorithm has a good effect in practical applications for the fast computation,high recognition and better recognition of skew, drift and information deformity etc. with the worse quality of fingerprints.