低分辨率人耳图像识别方法研究倡.pdf

文档介绍：第 27 卷第 11 .11
2010 年 11 月 Vol .2010No
Application Research puters Nov
低分辨率人耳图像识别方法研究倡
1,2 1 3
王晓云, 苑玮琦, 郭金玉
(, 沈阳 110870; , 沈阳 110168;
化工大学信息工程学院, 沈阳 110142)
摘要: 针对人耳识别中存储量和计算速度的要求,同时考虑远距离拍摄时低分辨率人耳识别问题,探讨了低
分辨率人耳图像识别性能,给出了分辨率与识别率的关系。首先采用高斯金字塔对人耳图像进行不同层的分
解,然后对每一层图像应用广义判别分析方法( )提取特征,最后计算样本间的余弦距离,通过阈值法分类识
GDA
别。实验结果表明,当人耳图像分辨率降低为 36 ×24 时系统识别性能最好,满足实时生物识别系统的要求。
关键词: 低分辨率; 人耳识别; ; 广义判别分析
Fisherear
中图分类号: 391畅41 文献标志码: 文章编号: 1001唱3695(2010)11唱4328唱03
TP A
: /. .
doi j issn
唱
Study of low resolution human ear image recognition
唱 1,2 , 唱 1 , 唱 3
WANG Xiao yun YUAN Wei qi GUO Jin yu
(puter Vision Group, Shenyang University of Technology, Shenyang 110780, China; of Mechanical Engineering, Shenyang
Ligong University, Shenyang 110168, China; of Information Engineering, Shenyang University of Chemical Technology, Shenyang
110142, China)
Abstract: 唱唱
This paper discussed the low resolution image recognition performance of the human ear and deduced the relation唱
ship between resolution and recognition唱 rate in view of the requirements for memory唱 capacity puting. speed in ear recog
nition and taking account of the low resolution ear recognition, problem at the long range shooting It first broke up the human唱
ear images into differen