基于局部分块特征的快速虹膜身份鉴别算法  

A fast iris identification algorithm based on local piece features

作　　者：苑玮琦[1] 张雷[1] 柯丽[1] 徐露[1] 

机构地区：[1]沈阳工业大学视觉检测技术研究所,辽宁沈阳110178

出　　处：《光电子．激光》2009年第5期648-652,共5页Journal of Optoelectronics·Laser

基　　金：国家自然科学基金资助项目(60672078;60472088)

摘　　要：提出一种新的虹膜身份鉴别算法。首先将灰度虹膜图像等分为若干个子图像,再将每个子图像等分为若干个子区域,将各点梯度之和最大的子区域中心点坐标作为各子图像的特征,各子图像的特征构成了该虹膜图像的特征矩阵,最后通过特征矩阵在空间上直接对准的方法进行匹配识别。给出了子图像和子区域大小的选择方法,同时给出了在人眼自然张开状态下不受遮挡干扰的可用虹膜区域大小对识别效果的影响,克服了之前人为规定子图像和子区域大小和选取固定可用虹膜区域大小提取局部纹理特征所带来的局限性。实验表明:1)人眼自然张开状态下,在虹膜区域超过50%不受遮挡时即可完成识别;2)算法运行速度快且对采集图像时左右各7°以内的旋转失真具有很强的容错能力。A new algorithm of iris identification is propose＆ First, the iris gray-scale image is divided into several sub images,and then these sub images are equally divided into several sub regions,The central coordinate of every sub region which contains the maximum gray scale sum of all points in this region will be considered as the feature coordinate of every sub image, And these coordinates will build the fea- ture matrix of the iris image. At last, the recognition is by matching the feature matrix in space directly. The paper describes the methods to find the sub images and sub regions as well as the influences of the size of usefuliris region, where there is no disturbance of sheltering, on the recognition when the eyes naturally open. It over courves the prerious limitation brought by artificially regulating the size of sub made and sub region and selecting a fixed size of extract the local the experimental results hayed proved that： 1） This method can perform recognition when there is only 50% or more iris region with no shelter when the eyes open naturally; 2） The algorithm has very fast operation speed and very high fault-toler- ant ability for the rotating distortion tess than 7°around caused by the image taken.

关 键 词：生物特征识别 虹膜识别 特征提取 匹配 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]