基于乳突纹方向特性的掌纹自动分类方法  

Automatic palmprint classification method based on the orientation of ridges

作　　者：张泽[1] 束为[1] 荣钢[1] 

机构地区：[1]清华大学自动化系,北京100084

出　　处：《清华大学学报（自然科学版）》2002年第9期1222-1224,共3页Journal of Tsinghua University(Science and Technology)

摘　　要：针对在利用掌纹进行身份自动鉴别的过程中如何对掌纹进行自动分类 ,提出了一种利用乳突纹方向特性的掌纹自动分类方法。此方法首先生成具有 4个方向的掌纹方向图 ,然后利用改进的庞加莱指数算法对掌纹方向图进行处理来确定三角点和中心点 ,最后根据掌纹指根区和掌纹外侧区中三角点和中心点的数量对掌纹进行分类。利用此方法对340个指根区和 35 4个外侧区的油墨捺印掌纹图像进行了测试 ,在指根区的正确率为 95 .6 % ,在外侧区则为 96 .9%。实验结果说明上述方法能够有效地进行掌纹自动分类。Palmprints can be automatically classified using an automatic palmprint classification method based on the orientation of ridges in the palmprint. The method generated a four -direction palmprint image which was then processed using the improved Poincaré index algorithm to determine the delta and core points. The number of delta and core points in the finger -root region and in the outside region were used to classify the palmprint. Tests with 340 inked palmprint images of the finger -root region and 354 inked palmprint images of the outside region had a correct identification rate of 95.6% for the finger region and 96.9% for the outside region. The experimental results show that the method is effective for automatic palmprint classification.

关 键 词：乳突纹 方向特性 掌纹自动分类 掌纹识别 掌纹方向图 身份鉴别 生物统计 图像处理 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] D918.91[自动化与计算机技术—计算机科学与技术]