一种有效的虹膜定位及睫毛检测方法  

Efficient algorithm on iris image localization and eyelash detection

作　　者：陈广华[1] 何希勤[1] 刘向东[1,2] 

机构地区：[1]辽宁科技大学应用数学研究所,辽宁鞍山114051 [2]大连民族学院非线性信息技术研究所,辽宁大连116600

出　　处：《计算机工程与应用》2010年第6期171-173,207,共4页Computer Engineering and Applications

基　　金：国家自然科学基金(No.60573124)~~

摘　　要：提出一种有效的虹膜定位及睫毛检测方法。通过把眼睛图像中分割成小的矩形区域,利用找到的这些矩形区域像素平均最小值把眼睛图像进行二值化,找到虹膜区域的内边界;以瞳孔的质心为参考点,在其左右的扇形区域内分别使用修改后的Daugman的检测算子,找到像素值变换大的位置,进而定位出虹膜的外界;使用Gobor滤波器和窗口移动法对睫毛进行有效的检测。通过对大量虹膜图像的实验表明,该方法取得了非常好的结果。An efficient algorithm on iris image localization and eyelash detection is presented.Firstly,an eye image is divided into many small rectangular regions.Image binarization is implemented taking the smallest average intensity of the regions as a threshold.The inner boundary of the iris region is found.Then the centroid of the pupil is taken as a reference point.In the left and right sector regions of the centroid,the modified Daugman＇s detecting operator is utilized to search the sharp variation position of the intensity.Then the outer boundary of the iris region is located.At last,eyelash is detected by using Gobor filter and the shift window.After doing a great deal of experiments,the results show that the proposed method has an encouraging performance.

关 键 词：虹膜识别 虹膜定位 睫毛检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]