摘要
提出一种基于多小波变换的新的虹膜识别算法.利用CL多小波优良的数学性质,虹膜纹理特征能够被准确完整地表达.将分解后子带作为特征编码,利用空间向量的方向夹角余弦作为匹配方法,有效克服了缩放、平移、光照等影响,达到了较好的识别效果.
This paper is proposed a new method for iris identification based on multiwavelet transform. By means of the mathematical properties of CL muhiwavelet, the iris texture feature can be presented accurately and primely. The sub-bands are coded and the information of direction cosine is used for iris matching, by means of which this method overcomes zoom, shift and illuminance and obtains approving effect.
出处
《吉林大学学报(理学版)》
CAS
CSCD
北大核心
2007年第1期68-72,共5页
Journal of Jilin University:Science Edition
基金
教育部高等学校博士点专项科研基金(批准号:20050183018)
吉林大学"985工程"科技创新基金
关键词
虹膜识别
多小波变换
纹理特征
iris identification
multiwavelet transform
texture feature