首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于语义的人眼特征快速自动提取方法
引用本文:刘祥楼.基于语义的人眼特征快速自动提取方法[J].科学技术与工程,2012,12(12):2849-2851,2855.
作者姓名:刘祥楼
作者单位:1. 东北石油大学;黑龙江省高校校企共建测试计量技术及仪器仪表工程研发中心,大庆163318
2. 东北石油大
基金项目:黑龙江省研究生创新科研项目(YJSCX2011-121HLJ)资助
摘    要:摘要:本文针对基于语义的人眼特征自动提取技术进行研究并探索出一条新的路径:首先,在人眼特征研究的基础上,自行建立了描述人眼特征的几何模型并定义了四个眼部特征参量;其次,采用三庭五眼规则分割出人眼区域;然后,利用梯度二值化及积分投影技术对人眼特征进行提取;最后,对人眼特征进行语义化并建立数字化标签。通过自建图库中1000幅彩色免冠照片的实验测试表明:计算机能够通过语义比对的方式实现快速人脸检索,最高重复率为8‰。

关 键 词:人眼特征提取  人脸检索  语义化
收稿时间:2/15/2012 1:56:12 PM
修稿时间:2/25/2012 8:43:47 AM

A Fast Method of Ocular Features Extraction Based on the Semantics
liuxianglou.A Fast Method of Ocular Features Extraction Based on the Semantics[J].Science Technology and Engineering,2012,12(12):2849-2851,2855.
Authors:liuxianglou
Institution:1(Northeast Petroleum University1,Daqing 163318,P.R.China;The University-enterprise R & D Center of Measuring and Testing Technology & Instrument and Meter Engineering in Heilongjiang Province2,Daqing 163318,P.R.China)
Abstract:The study of ocular features extraction technology based on the semantics and discover a new path is expressed.First of all,on the basis of ocular features study,the geometric model was established for describe the ocular features and the four parameters were defined.Secondly,the eyes boundaries were separated by adopt three atriums and five eyes rule.Then,use of the grads binaryzation and integral projection technique for extraction the ocular features is made.Finally,the ocular features semantization and digitization label establishment.Through experiment and testing 1 000 full-faced colour photos in self-created photo gallery,the results indicate that: computer is able to bring about the fast facial image retrieval by pass the way of semantic comparison,maximum repetition rate 0.8%
Keywords:ocular features extraction  facial image retrieval  semantization
本文献已被 CNKI 万方数据 等数据库收录!
点击此处可从《科学技术与工程》浏览原始摘要信息
点击此处可从《科学技术与工程》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号