虹膜区域预处理及纹理特征提取方法的研究

摘要

随着电子信息技术的发展,个人的一些隐私信息越来越容易受到侵犯,而基于虹膜纹理特征的虹膜识别技术在确保个人信息安全这一块扮演着越来越重要的角色.在结构上分为预处理和纹理特征提取两部分.在预处理部分,提出一种基于共形几何代数理论结合最小二乘拟合的虹膜区域边界定位方法;在虹膜纹理特征提取部分,提出一种基于多特征的纹理提取方法.实验结果证明,在虹膜区域定位这一块,该方法定位结果准确,能实现对虹膜区域内外边缘的同时定位;在纹理特征提取部分,基于多特征的纹理图像查准率有了很大的提高.%With the development of electronic information technology,the individual privacy information is increasingly vulnerable,and the iris recognition technology based on iris texture feature plays an important role in ensuring the individual information security.The technology is structurally composed of the preprocessing section and texture feature extraction section.An iris region boundary localization method based on conformal geometric algebra (CGA) theory and least square fitting is proposed in the first paragraph of this paper.A texture extraction method based on multi-feature is proposed in the second paragraph of this paper.The experimental results show that the method has accurate positioning result,and can locate the inner and outer edges of iris region simultaneously in the aspect of iris region localization;its texture image precision ratio based on multi-feature is improved greatly in the aspect of texture feature extraction.