一种基于结构特征的虹膜识别方法  

A New Method for Iris Recognition

作　　者：林金龙[1] 石青云[1] 

机构地区：[1]北京大学信息科学中心,北京100871

出　　处：《计算机工程与应用》2003年第13期83-84,137,共3页Computer Engineering and Applications

摘　　要：提出了一种基于特征点比对的虹膜识别方法。与现有利用纹理分析进行识别的方法相比,这一方法具有编码效率高,抗噪声能力强,识别准确性好等特点。该方法首先利用LOG(LaplacianofGaussian)滤波器提取特征区域,然后在特征区中确定特征点。在方法中提出了基于特征点的相似性度量方法和弹性特征点比对方法。试验结果表明:这种方法分类效果较好。In this paper,a new method for iris recognition is presented.Compared with most different form of methods at present ,which are mainly based on texture analysis,this method use physical structure characteristics to classify iris images.Laplacian of Gaussian is used to extract minutiae features and the total distance of minutiae in two irises is used to represent the comparability of irises.Compared with existing methods the new method is less relative with image intensity and more stable.

关 键 词：结构特征 虹膜识别 身份识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]