摘要
针对传统虹膜定位算法定位速度慢、受眼睑、睫毛等遮挡影响的不足,本文基于投票竞选和圆的几何特性,提出基于预优投票竞选策略的虹膜圆检测算法。为减少盲目搜索和克服眼睑、睫毛等遮挡影响,建立了逐步求精的实施方案,从而形成一个快速有效、鲁棒性强的虹膜定位算法。基于JLU-IRIS虹膜库,从定位速度和定位精度两方面评价本文构造算法性能,将本文构造算法与积分微分算子法、Hough圆检测算法进行了对比测试。实验结果表明,本文方法具有快速、鲁棒性强的特点。
Traditional iris location algorithms that have drawbacks of slow locating and partial occlusion influences such as eyelids,eyelashes and so on,the paper presented a preconditioning voting circle detection algorithm based on preconditioning voting and circle geometry property.In additional,in order to reduce the computation and effect by partial occlusion,such as eyelid,eyelashes etc,the iris boundaries are localized,using a coarse-to-fine strategy,in regions of interest.In experiments,the proposed method is compared with the integral-differential operator and Hough transform algorithm based on JLU-IRIS iris database.From the experiment results,the proposed method is faster and better than the other two methods.
出处
《吉林大学学报（工学版）》
EI
CAS
CSCD
北大核心
2009年第S1期254-259,共6页
Journal of Jilin University:Engineering and Technology Edition
基金
国家自然科学基金项目(50422285)
高等学校博士学科点专项科研基金项目(20050183018)
吉林大学"985工程"项目
关键词
计算机应用
虹膜识别
虹膜定位
圆检测算法
预优投票
computer application
iris recognition
iris location
circle detection
preconditioning voting
