摘要
典型的虹膜识别算法都是针对单个虹膜样本采用图像处理的方法提取虹膜特征,只强调个体信息,而忽略了虹膜样本间的联系.本文提出了一种基于普通向量的虹膜识别算法,从模式分析的角度,利用普通向量方法融合虹膜的个体特征和统计特征,并且通过高效的睫毛滤除和眼皮定位的算法,进一步提高了识别的精度和算法的稳定性.实验结果表明了该算法具有良好的性能.
TypiCal iris recognition algorithms extract single iris sampie's feature via image processing methods which only emphasize on individual information without taking the relationship among iris samples into consideration. The paper proposes a novel iris recognition algorithm based on common vector. The algorithm combines individual features with statistical features by using the common vector method,. Additionally high efficient eyelashes removing and eyelids locating algorithms make further improve- ment to the precision and stability of the algorithm. The experimental results show the high perfomamce of the algorithm.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2007年第B12期69-73,共5页
Acta Electronica Sinica
基金
国家自然科学基金(No.60671064,No.60703011)
全国百名优秀博士学位论文作者专项资金(No.FANEDD-200238)
国家863计划(N02007AA012458)
教育部新世纪优秀人才支持计划(No.NCET-04-0330)
高等学校博士学科专项科研基金(No.RFDP-20070213047)
关键词
虹膜识别
个体特征
统计特征
普通向量
iris recognition
individual feature
statistical feature
common vector