基于SUSAN的指纹细节点提取算法  

An Algorithm for Extraction of Fingerprint Minutiae Based on SUSAN

作　　者：文潇[1] 华骅[2] 徐安士[1] 

机构地区：[1]北京大学区域光通信网与新型光通信系统国家重点实验室,北京100871 [2]四川大学电子信息学院,成都610064

出　　处：《四川大学学报（自然科学版）》2005年第5期952-956,共5页Journal of Sichuan University(Natural Science Edition)

基　　金：国家自然科学基金(60272023);教育部基金(90104003)

摘　　要：在整个指纹识别过程中,指纹图像的细节点提取至关重要.许多算法是先将脊线细化后进行提取,耗时较多,且易产生大量不易去除的伪细节点.作者将一般图象配准中常用的角点检测技术运用到指纹识别中,提出基于SUSAN的指纹细节点提取新算法.该法在传统的SUSAN上进行数项改进,使其适用于指纹图象,完成端点和分叉点的提取,再根据所产生伪细节点的分布情况将它们去除.此法计算简单、抗噪声能力强、不需事先对脊线进行细化.Minutiae extraction plays an important role in fingerprint identification. Most of the algorithms require the ridges being thinning first, which is time consuming and would produce a lot of fake minutiae. A new algorithm based on SUSAN in corner detection is proposed for fingerprint identification. It makes several ameliorations on traditional SUSAN to serve the fingerprint turn, uses the algorithm to extract both of the endings and bifurcations, and reduce the number of fake minutiae according to their distribution. This algorithm is simple, noise- resisting, and ridge thinning free.

关 键 词：端点 分叉点 角点检测 SUSAN 伪细节点 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]