基于测地距离的KPCA人脸识别

计算机工程与科学

基于测地距离的KPCA人脸识别

林克正，魏颖，钟岩，李慧

（哈尔滨理工大学计算机科学与技术学院，黑龙江哈尔滨 150080）

收稿日期:2015-08-17 修回日期:2015-09-30 出版日期:2016-09-25 发布日期:2016-09-25
基金资助:
国家自然科学基金(60873019);黑龙江省教育厅科学技术研究项目(11551087)

KPCA face recognition based on geodesic distance

LIN Ke-zheng,WEI Ying,ZHONG Yan,LI Hui

（School of Computer Science and Technology,Harbin University of Science and Technology,Harbin 150080,China）

Received:2015-08-17 Revised:2015-09-30 Online:2016-09-25 Published:2016-09-25

摘要/Abstract

摘要：

针对人脸检测数据集中的信息均为高维特征向量且人脸识别易受表情变化影响等问题，本文提出一种基于测地距离的KPCA人脸识别方法，该方法利用非线性方法提取主成分。先采用KPCA方法把人脸数据映射到高维空间，进而在高维空间中提取人脸的主成分，其中核函数为多项式核函数；然后引入测地距离替换原来的欧氏距离进行相似度量，其能更准确地测量出两像素点间的实际距离，使得人脸识别率受表情变化影响小。该方法不但可以实现降维，而且还能达到有效提取特征的目的。在ORL人脸库上的实验结果表明，该方法的识别率明显优于PCA、KPCA等方法的识别率。

关键词: 人脸识别, 特征提取, 主成分分析, 核函数, 测地距离

Abstract:

Aiming at the problem that the information of the face detection data is high eigenvector and face recognition is easily affected by expression changes, we propose a kernel principle component analysis (KPCA) face recognition method based on geodesic distance. We extract principal components by the nonlinear method. First, we adopt the KPCA method to map face data to the high dimensional space, and then principal components of the face are extracted in the high dimensional space, where the kernel function is a polynomial kernel. Finally, geodesic distance is introduced to replace the original Euclidean distance as the similarity measure, which can more accurately measure the actual distance between two points, and face recognition rate is less affected by expression changes as well. The proposed method cannot only achieve dimension reduction, but also achieve effective data extraction. Experiments on the ORL face database show that the recognition rate of the proposed method is superior to that of the PCA and the KPCA.

Key words: face recognition, feature extraction, PCA, kernel function, geodesic distance

林克正，魏颖，钟岩，李慧. 基于测地距离的KPCA人脸识别[J]. 计算机工程与科学.

LIN Ke-zheng,WEI Ying,ZHONG Yan,LI Hui. KPCA face recognition based on geodesic distance [J]. Computer Engineering & Science.

参考文献

［1］Huang Jin-yu,Zhang Hui-lin,Yan Ri-liang.Face recognition of LBP histogram PCA and education distance ［J］.Computer Systems & Applications,2012,21(6):201-205.(in Chinese)
［2］Liu Li,Kuang Gang-yao.Overview of image textural feature extraction methods ［J］.Journal of Image and Graphics,2009,14(4):622-635.(in Chinese)
［3］Adalarasan R,Santhanakumar M,Sundaram A S.Optimization of weld characteristics of friction welded AA 6061-AA 6351 joints using grey-principal component analysis (G-PCA) ［J］.Journal of Mechanical Science and Technology,2014,28(1):301-307.
［4］Wang Hua-zhong,Yu Jin-shou.Study on the kernel-based methods and its models selection ［J］.Journal of Southern Yangtze University(Natural Science Edition),2006,5(4):500-504.(in Chinese)
［5］Zhao Jian-hua,Wang Shun-fang,Zhang Fei-long.Face recognition study with combination-kernel-based KPCA ［J］.Computer Engineering and Design,2014,35(2):631-635.(in Chinese)
［6］Hofmann T,Scholkopf B,Smols A J.Kernel methods in machine learning ［J］.Annals of Statistics,2008,36(3):1171-1120.
［7］Guo Di,Zhou Ming-quan,Wang Xue-song,et al.Simulation of blemishes appearance of three-dimensional models based on geodesic distance ［J］.Journal of System Simulation,2014,26(9):1984-1988.(in Chinese)
［8］Sonnenburg S,Ratsch G,Schafer C,et al.Large scale multiple kernel learning ［J］.Journal of Machine Learning Research,2006,7:1531-1565.
［9］Atonen T,Hadid A,Pietikainen M.Face description with local binary patterns:Application to face recognition ［J］.IEEE Transactions on PAMI,2006,28(12):2037-2041.
［10］Wu Wen-tong, Li Yuan-xiang, Wei Bang-he, et al. Incremental ISOMAP method based on locally estimated geodesic distance ［J］.Journal of Shanghai Jiaotong University,2013,47(7):1082-1086.(in Chinese)
［11］Jiang Yan-huang,Yang Xue-jun.Research on stacked classifiers ［J］.Computer Engineering & Science,2004,26(6):67-76.(in Chinese)
［12］Gao Shan-shan,Chi Jing,Liu Cui-yun,et al.Noisy image segmentation based on approximate geodesic distance ［J］.Journal of Computer-Aided Design & Computer Graphics,2014,26(12):2214-2222.(in Chinese)
［13］Lü Ning,Liu Shao-bo,Yu Xiao-yang.A compound KPCA fault diagnosis model ［J］.Journal of North University of China,2009,30(6):555-560.(in Chinese)
［14］Liang Li-ming,Feng Xin-gang,Chen Yun-nen,et al.Method of selection kernel function based on distribution characteristics of samples ［J］.Computer Simulation,2013,30(1):523-528.(in Chinese)
附中文参考文献：
［1］黄金钰,张会林,闫日亮.LBP直方图与PCA的欧氏距离的人脸识别［J］.计算机系统应用,2012,21(6):201-205.
［2］刘丽,匡纲要.图像纹理特征提取方法综述［J］.中国图象图形学报,2009,14(4):622-635.
［4］王华忠,俞金寿.核函数方法及其模型选择［J］.江南大学学报(自然科学版),2006,5(4):500-504.
［5］赵剑华,王顺芳,张飞龙.基于组合核函数KPCA的人脸识别研究［J］.计算机工程与设计,2014,35(2):631-635.
［7］郭迪,周明全,王学松,等.基于测地距离三维模型瑕疵外观模拟［J］.系统仿真学报,2014,26(9):1984-1988.
［10］吴文通，李元祥，韦邦合，等.局部测地距离估计的增量等距特征映射算法［J］.上海交通大学学报，2013，47（7）：1082-1086.
［11］蒋艳凰,杨学军.多层组合分类器研究［J］.计算机工程与科学,2004,26(6):67-76.
［12］高珊珊,迟静,刘翠云,等.近似测地距离度量下的图像抗噪分割方法［J］.计算机辅助设计与图形学学报,2014,26(12):2214-2222.
［13］吕宁,刘少波,于晓洋.一种复合KPCA故障诊断模型［J］.中北大学学报,2009,30(6):555-560.
［14］梁礼明,冯新刚,陈云嫩,等.基于样本分布特征的核函数选择方法研究［J］.计算机仿真,2013,30(1):523-528.

[1]	张建伟, 周亚同, 史宝军, 何昊, 王文. 基于SSD算法的轻量化仪器表盘检测算法[J]. 计算机工程与科学, 2022, 44(08): 1418-1425.
[2]	刘云, 郑文凤, 张轶. 代价约束算法对入侵检测特征提取的优化研究[J]. 计算机工程与科学, 2022, 44(03): 447-453.
[3]	严春满, 张昱瑶, 张迪. 改进标签一致KSVD字典学习的人脸识别算法[J]. 计算机工程与科学, 2022, 44(02): 291-297.
[4]	谈恩民, 王晨. 基于表征学习的模拟电路故障诊断[J]. 计算机工程与科学, 2022, 44(01): 27-35.
[5]	姬子恒, 王斌. 基于深度学习的草图检索方法研究进展[J]. 计算机工程与科学, 2021, 43(12): 2190-2205.
[6]	王鑫, 廖祎玮, 赵国生, 王健, 谢宝文. 一种面向任务需求的群智感知任务分配模型[J]. 计算机工程与科学, 2021, 43(08): 1512-1520.
[7]	党小超, 汝春瑞, 郝占军, . 基于CSI与SVM回归的室内定位方法[J]. 计算机工程与科学, 2021, 43(05): 853-861.
[8]	李晶晶, 许少华. 基于组合核函数的径向基过程神经网络及其在示功图诊断中的应用[J]. 计算机工程与科学, 2021, 43(04): 746-752.
[9]	孙雨浩, 陶洋, 胡昊. 基于判别低秩矩阵恢复和协同表示的遮挡人脸识别[J]. 计算机工程与科学, 2020, 42(12): 2199-2207.
[10]	张永梅, 滑瑞敏, 马健喆, 胡蕾. 基于深度学习与超分辨率重建的遥感高时空融合方法[J]. 计算机工程与科学, 2020, 42(09): 1578-1586.
[11]	王思齐, 胡婧韬, 余广, 祝恩, 蔡志平. 智能视频异常事件检测方法综述[J]. 计算机工程与科学, 2020, 42(08): 1393-1405.
[12]	沈晓燕, 王雪梅, 王燕. 基于样本熵和模式识别的脑电信号识别算法研究[J]. 计算机工程与科学, 2020, 42(08): 1482-1488.
[13]	盛明伟，唐松奇，万磊，秦洪德，李俊. 基于改进CNN-RANSAC的水下图像特征配准方法[J]. 计算机工程与科学, 2020, 42(05): 859-868.
[14]	付文博1，2，何欣1，2，于俊洋1，2 . C-Canny算法和改进单层神经网络相结合的面部特征点定位[J]. 计算机工程与科学, 2020, 42(04): 658-664.
[15]	钟国韵,王檬檬,汪宇玲,常艳荣,吴忠粱. 旋转均值跳动特征及其在模糊人脸识别中的应用[J]. 计算机工程与科学, 2020, 42(03): 474-482.