一种新的基于结构特征的虹膜识别算法  

A Novel Iris Recognition Method Based on the Structures and Textures of Irises

作　　者：苑玮琦[1] 徐露[1] 林忠华[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110023

出　　处：《光电子．激光》2006年第10期1250-1254,共5页Journal of Optoelectronics·Laser

基　　金：国家自然科学基金资助项目(60472088);辽宁省自然科学基金资助项目(20042028)

摘　　要：提出了一种新的基于结构特征的虹膜识别算法。其基本思想是:首先,通过图像预处理得到归一化的虹膜图像;其次,利用逻辑判断的方法在虹膜图像上寻找局部极小值点;再次,去掉伪特征点,生成固定长度的虹膜编码;最后,计算2个虹膜编码间的“与”距离,根据该距离给出识别结果。在该CASIA虹膜图像数据库上进行了大量的实验,结果表明,该算法与现有的虹膜识别算法相比,识别准确率更高,识别速度更快。A novel iris recognition method is described based on the structures of irises. Firstly, we get the fixed-size iris image through the image preprocessing. Secondly, we detect all of the feature points in the iris image utilizing the method of finding local minimum points. Thirdly,we abandon the fake feature points and create the iris codes. Finally,we calculate the Hamming distance between two iris codes and obtain the recognition result according to the Hamming distance. The experimental results on 756 iris images show that the accuracy of the proposed method is higher than other recognition algorithms found in the current literatures, and the recognition speed is very fast.

关 键 词：生物特征 虹膜识别 特征提取 虹膜编码 匹配 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]