手背静脉特征提取算法  

Recognition Algorithm of Dorsal Hand Vein

作　　者：韩笑[1] 冀进朝[2] 左平[3] 张禹[2] 

机构地区：[1]吉林大学数学学院,长春130012 [2]吉林大学计算机科学与技术学院,长春130012 [3]空军航空大学基础部,长春130022

出　　处：《吉林大学学报（理学版）》2010年第2期245-250,共6页Journal of Jilin University:Science Edition

基　　金：国家自然科学基金(批准号:10926157)

摘　　要：提出一种基于有限Radon变换(FRAT)的手背静脉识别新方法.在考虑手背静脉图像"线状"特点的基础上,先将手背静脉图像"分块",再对分块后的图像进行FRAT变换,并提出一种向量非均衡分布可得最大值的方法,提取手背静脉图像纹理特征,最后通过特征匹配进行分类识别.结果表明:与传统的Hough变换和离散Radon变换相比,新方法提高了目标直线特征的检测速度;在有800个样本的数据库上进行实验,获得了96.5%的识别率,表明了算法的有效性.A new method for dorsal hand vein recognition based on FRAT was proposed. We firstly partitioned these hand vein images into blocks on the basis of their ＂line＂ features ; then did FRAT on the block images, and put up a new method to get the maximum by means of the non-equilibrium distribution of vector in order to get the texture features; finally, obtained classification results through feature matching. Compared with discrete Radon transform and traditional Hough transform, our method improves the detection speed of straight line features. At the same time, our method was applied to the test of 800 samples, showing a recognition rate of 96.5%. It has demonstrated that this algorithm is effective.

关 键 词：有限Radon变换(FRAT) 手背静脉 特征提取算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]