基于多通道Gabor滤波与灰度梯度共生矩阵的虹膜识别方法  

Iris Recognition Method Based on Multi-channel Gabor Filtering and GGCM

作　　者：白雪冰[1] 张庭亮[1] 祝贺[1] 

机构地区：[1]东北林业大学机电工程学院,哈尔滨150040

出　　处：《现代科学仪器》2013年第5期69-72,94,共5页Modern Scientific Instruments

基　　金：黑龙江省自然科学基金项目(C201208);黑龙江省博士后基金项目(LBH-Q10160)

摘　　要：为有效改善虹膜识别系统的性能,针对传统方法提取全局特征时对局部纹理特征不敏感的问题,提出了一种基于多通道Gabor滤波和灰度梯度共生矩阵(GGCM)的虹膜识别方法。使用不同方向和尺度的Gabor滤波器组对预处理后的虹膜图像滤波,由全局滤波图像构建灰度梯度共生矩阵,提取特征值生成虹膜特向量。识别过程采用加权欧氏距离进行特征匹配。CASIA虹膜图像库实验表明,该方法能在保证识别系统实时性能的前提下使EER下降0.5%。In order to improve the performance of iris recognition system,aiming at the problem that local texture is ignored when it comes to the traditional method to extract the global features,an iris recognition method based on multi- channel Gabor filtering and GGCM is proposed.First of all,the feature image was generated through different scales and directions of Gabor filter, then GGCM was built by the global filtered image,and feature value was extracted to generate the iris feature vector.The recognition process used weighted Euclidean distance to match with it.Experiments on CASIA iris showed that this method can guarantee real-time performance of the identification system under the premise of EER down 0.5%.

关 键 词：虹膜识别 GABOR滤波 灰度梯度共生矩阵 加权欧式距离 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]