正交Haar变换的眉毛识别方法*
Orthogonal Haar Transform for Eyebrow Recognition

 

李厚君,王日凤,李春贵
LI Houjun,WANG Rifeng,LI Chungui
 

(广西科技大学计算机科学与通信工程学院,广西柳州   545006)
(School of Computer Science and Communication Engineering,Guangxi University of Science and Technology,Liuzhou,Guangxi,545006,China)


摘要:【目的】提高现有眉毛识别方法的识别效率。【方法】采用快速正交Haar变换模板匹配算法(FOHT),设计一种基于正交Haar变换的眉毛识别方法;同时,使用最大标准子模板和自适应阈值解决了FOHT算法只能处理标准模板且需要手动设置阈值的缺陷。【结果】所构建的眉毛识别方法比原方法效率提高约32%。【结论】正交Haar变换的眉毛识别方法实时性强,具有一定的实用价值。
关键词:眉毛识别   正交Haar变换   模板匹配   最大标准子模板   自适应阈值
中图分类号:TP181      文献标识码:A       文章编号:1002-7378(2016)01-0036-06
Abstract:【Objective】To improve the efficiency of current method on the eyebrow recognition.【Methods】Orthogonal Haar Transform Eyebrow Recognition method was established by using fast orthogonal Haar Transform template matching algorithm (FOHT).The approaches of maximum standard template and adaptive threshold were used to solve the problems of FOHT,which could only process the situation of standard template and need to set the threshold manually.These methods increased the automatic degree of eyebrow recognition,and improved the recognition efficiency.【Results】Experimental results showed that the recognition efficiency of the proposed method increased about 32% than that of the original method.【Conclusion】The orthogonal Haar transform eyebrow recognition method shows better realtime performance and is practically valuable for the promotion and application of eyebrow recognition.
Key words:eyebrow recognition,orthogonal Haar transform,template matching,maximum standard template,adaptive threshold

 

*国家自然科学基金项目(61440017),广西科技大学博士基金项目(院科博11z13)和广西多源信息挖掘与安全重点实验室开放基金项目(MIMS13-04)资助。

 

PDF全文下载cbimg.cnki.net/Editor/2016/0214/gxkx/c439056e-2d7e-46ae-9228-babada2f8605.pdf

发布日期:2016/2/14