基于2DPCA的眉毛识别方法研究

基于2DPCA的眉毛识别方法研究 Study of Eyebrow Recognition Based on 2DPCA 李玉鑑 LI Yujian 1 first-author 谢欢曦 XIE Huanxi 1 周艺华 ZHOU Yihua 1 北京工业大学计算机学院 北京工业大学计算机学院 College of Computer Science,Beijing University of Technology College of Computer Science,Beijing University of Technology 分析了人类的眉毛作为一种独立生物特征进行身份鉴别的可能性,提出了一种基于2DPCA的眉毛识别方法,比较了2DPCA和PCA对眉毛识别的结果.通过对109人的眉毛数据库的分析,分别研究了2DPCA和PCA方法在眉毛识别中识别率和信息阈值之间,识别率与特征个数之间,识别率与图像尺寸之间的关系,以及它们对时间和空间的需求.实验结果表明,利用2DPCA方法进行眉毛识别在训练时间和识别率上明显优于PCA方法,识别率最高可达97.62%. This paper validates that it is feasible to use eyebrow as an independent biometric in personal identification and then we propose an eyebrow recognition method based on 2DPCA.Moreover,we compare the influence of 2DPCA and PCA on eyebrow recognition.By the eyebrow database taken from 109 persons,we analyze the relationship of their eyebrow recognition accuracies with information threshold,the number of features,the size of image respectively as well as their requirement for time and space.The results show that the eyebrow recognition using 2DPCA,with the highest accuracy of 97.62%,is obviously superior to using PCA in terms of training time and recognition accuracy. 生物特征识别 眉毛识别 2DPCA 特征提取 biometric recognition eyebrow recognition two-dimensional principle component analysis（2DPCA） feature extraction 10.14188/j.1671-8836.2011.06.010 TP391.41 国家自然科学基金（60775010）;北京市自然科学基金资助项目（4112009,4113067）;北京市教委科技发展项目（KM201010005012）;北京工业大学高层次人才培养项目资助 2011-06-01 2021-04-01 6 57