摘要
为了提高虹膜识别的准确率,通过对虹膜图像进行处理,实现了对虹膜图像的准确定位,得到了增强的归一化图像;使用Haar小波变换进行了特征提取,通过采用K-means方法对小波特征数据进行聚类,实现了粗分类得到了小样本集虹膜图像;结合虹膜的纹理特点,通过使用Log-Gabor滤波器提取虹膜局部纹理特征,量化编码后形成了虹膜特征模板;然后在得到的小样本集内通过汉明距离计算虹膜特征模板的相似度,完成对虹膜图像的识别。实验结果表明,提出的虹膜识别方法有效地避免了虹膜匹配过程中因为虹膜数据库中种类多、数量多带来的计算量大、计算时间长的问题,提高了识别准确率。
In order to improve recognition accuracy of iris recognition,the iris image was disposed,as the result the iris region of image was accurately located and the normalized image was enhanced.The Haar wavelet transform was used in the feature extrac-tion and K-means was used to cluster the feature,so that a small sample set of iris images was obtained.Combined with iris texture characteristics,the features were extracted by using Log-Gabor filter and the iris feature template was formed after quantization cod-ing.The similarity of iris feature template was calculated by Hamming distance in small sample set,and the iris recognition was completed.The testing results illustrate that the proposed algorithm has a certain improvement in recognition accuracy and effectively avoids the problem of large amount of computation and long time in iris matching because of the variety and quantity of iris database.
作者
姚立平
潘中良
Yao Liping;Pan Zhongliang(College of Physics and Telecommunications Engineering,South China Normal University,Guangzhou 510006,China)
出处
《电子技术应用》
2019年第4期113-117,共5页
Application of Electronic Technique
