摘要
本文提出了一种改进的基于人眼结构特征的虹膜识别方法,基本思想是:首先,通过图像预处理,确定虹膜的位置和大小,将环形虹膜图像展开成矩形并进行归一化与图像增强;其次,利用局部灰度极小值的方法寻找有效虹膜区域内的特征点,依据纹理长度和方向信息去掉伪特征点,得到虹膜二进制编码;最后根据匹配准则进行识别。大量实验表明,该方法识别率高,识别时间短。
Because of the unique structure characteristics and abundant information of irises, iris recognition has the highest accuracy in all of the personal identification technologies. An improved iris recognition algorithm is presented based on the structures of human eyes. Firstly, get the location and size of the iris through the image preprocessing, and then unwrap the localized iris to a rectangular texture block and make normalization and enhancement. Secondly, detect all of the feature points in the effective iris area utilizing the method of local gray minimum, and then abandon the false feature points by the information of texture length and direction and get a binary iris code, Finally, obtain the recognition result according to the matching rules.
出处
《光电工程》
EI
CAS
CSCD
北大核心
2007年第8期105-109,133,共6页
Opto-Electronic Engineering
基金
国家自然科学基金资助项目(60672078
60472088)
辽宁省自然科学基金资助项目(20042028)
关键词
虹膜识别
图像预处理
特征提取
图像编码
iris recognition
image preprocessing
feature extraction
image encoding
