应用PCA理论进行多人脸姿态估计的方法

摘要/Abstract

摘要：

多人脸姿态估计是多姿态人脸识别的关键问题,至今尚未很好地解决。本文提出了一种基于PCA(Principal Component Analysis)算法的多人脸姿态估计方法,以获得人脸姿态与人脸图像在特征空间投影间的对应关系。算法首先利用PCA理论构建训练样本集的姿态子空间与特征向量,然后利用欧氏投影距离进行姿态估计,最后对方差贡献率与姿态估计准确率间的关系进行了研究。为了验证算法的有效性,利用23类姿态,共690个样本进行了实验,实验结果表明,该方法的姿态估计准确率为84%。说明将该方法应用与多姿态人脸估计是有效的,可行的。

关键词: 人脸姿态估计, 主成分分析, 姿态子空间, 方差贡献率, 特征向量

Abstract:

Pose estimation of different human faces is a key technology of varied human face recognitions which has not been well resolved yet.In order to get the mapping between face poses and feature spaces,an algorithm based on PCA was presented to realize the face pose estimations.Firstly,the subspaces and eigenvectors of different poses was built and selected by using PCA theory.Secondly,the Euclidean distance classifier was chosen to estimate the multi-poses of human faces.Lastly,the relationship between PCA's Variance contribution rates and pose estimation accuracy rates was investigated in detail.The algorithm was validated by means of experimental tests of human face pose database,which contains 23 kinds of poses and 630 samples in detail;experimental results show that the precision of this algorithm is above 84%.It also shows that the PCA based pose estimation method is feasible and valid for varied human face recognition.

Key words: face pose estimation, principal component analysis(PCA), pose subspace, variance contribution, eigenvectors

中图分类号:

TP391.41

宋怀波, 史建强. 应用PCA理论进行多人脸姿态估计的方法[J]. 吉林大学学报(工学版), 2013, 43(增刊1): 43-46.

SONG Huai-bo, SHI Jian-qiang. Pose estimation of varied human faces based on PCA method[J]. 吉林大学学报(工学版), 2013, 43(增刊1): 43-46.

参考文献

[1] Ji Qiang,Hu Rong.3D face pose estimation and tracking from a monocular camera [J].Image and Vision Computing,2002,20(7):499-511.

[2] 薛雨丽,毛峡,郭叶,等.人脸交互中的人脸表情识别研究进展[J].中国图象图形学报,2009,14(5):764-772. Xue Yu-li,Mao Xia,Guo Ye,et al.The research advance of facial expression recognition in human computer interaction [J].Journal of image and graphics,2009,14(5):746-772.

[3] 宋红,石峰.基于人脸检测与跟踪的智能监控系统[J].北京理工大学学报,2004,24(11):966-970. Song H,Shi F.Face detection and tracking for intelligent surveillance system [J].Journal of Beijing Institute of Technology,2004,24(11):966- 970.

[4] 李旭东.用于人脸表情图像与动画合成的图像变形方法[J].武汉大学学报信息科学版,2007,32(9):796- 799. Li Xu-dong.Image morphing method for facial expression image and animation synthesis [J].Journal of Geomatics and Information Science of Wuhan University,2007,32(9):796-799.

[5] 李志春.驾驶员疲劳状态检测技术研究与工程实现.镇江:江苏大学,2009. Li Zhi-chun.Research on driver's fatigue status detection technology and realization in engineering .Zhenjiang:Jiangsu University,2009.

[6] Yun Fu,Thomas Huang.Graph embedded analysis for head pose estimation[C]//7th International Conference on Automatic Face and Gesture Recognition,2006:3-8.

[7] 刘金葵.正面人脸判别的方法研究[D].广州:中山大学,2009. Liu Jin-kui.The research of the positive face discrimination [D].Guangzhou:Central South University,2009.

[8] 孙昊翔.关于人脸姿态估计方法的研究[D].长春:吉林大学,2007. Sun Hao-xiang.The research of the face pose estimation[D].Changchun:Jilin University,2007.

[9] Chen Q,Wu H,Shimada T.A robust algorithm for 3D head pose estimation[C]// International Conference on Multimedia Computing and Systems,Nara,1999:697-702.

[10] Hu Y X,Chen L B,Zhou Y,et a1.Estimating face pose by facial asymmetry and geometry[C]// IEEE Conf.on FG'04,2004:651-656.

[11] Hu F,Wang Y,Wu H.Mobile telemedicine sensor net-works with low-energy data query and network lifetime considerations [J].Mobile Computing,2007,5(4):404-417.

[12] 张媛,张燕平.一种PCA算法及其应用[J].微机发展.2005,15(2):69-72. Zhang Yuan,Zhang Yan-ping.An algorithm of PCA and its application [J].Microcomputer Development,2005,15(2):69- 72.

[13] 王宏拥,廖海斌,段新华,等.基于奇异值与特征融合矩阵的自适应人脸识别[J].计算机工程与应用,2010,46(7):162-164. Wang H Y,Liao H B,Duan X H,et al.Weighted adaptive face recognition using singular value and feature-matrix[J].Computer Engineering and Applications,2010,46(7):162-164.

[14] 孙即祥.现代模式识别[M].北京:国防科技大学出版社,2002.

相关文章 15

[1]	刘哲, 徐涛, 宋余庆, 徐春艳. 基于NSCT变换和相似信息鲁棒主成分分析模型的图像融合技术[J]. 吉林大学学报(工学版), 2018, 48(5): 1614-1620.
[2]	张曼, 施树明. 典型汽车运行工况的状态转移特征分析[J]. 吉林大学学报(工学版), 2018, 48(4): 1008-1015.
[3]	耿庆田, 于繁华, 王宇婷, 高琦坤. 基于特征融合的车型检测新算法[J]. 吉林大学学报(工学版), 2018, 48(3): 929-935.
[4]	万成彪, 潘孟春, 张琦, 庞鸿锋, 朱学军. 基于张量特征值和特征向量的磁性目标定位[J]. 吉林大学学报(工学版), 2017, 47(2): 655-660.
[5]	苏畅, 付黎明, 魏君, 李硕, 黄蕾, 曹越. 基于感性工学和主成分分析的车身色彩设计[J]. 吉林大学学报(工学版), 2016, 46(5): 1414-1419.
[6]	申铉京, 范子龙, 吕颖达, 陈海鹏. 基于统计特征的图像篡改评价模型[J]. 吉林大学学报(工学版), 2016, 46(4): 1232-1238.
[7]	马爽, 周长久, 张连东, 洪伟, 田彦涛. 基于Kinect改进的增量PCA扭锁在线识别[J]. 吉林大学学报(工学版), 2016, 46(3): 890-896.
[8]	付永庆, 刘伟. 基于时间反转算子分解的选择性聚焦方法[J]. 吉林大学学报(工学版), 2015, 45(5): 1731-1736.
[9]	齐兴达, 李显君, 刘丝雨, 孟东晖. 基于数据包络分析和主成分分析的产业技术创新能力差异化研究[J]. 吉林大学学报(工学版), 2015, 45(3): 1017-1023.
[10]	姜万录，卢传奇，朱勇. 基于HHT和模糊C均值聚类的轴向柱塞泵故障识别[J]. 吉林大学学报(工学版), 2015, 45(2): 429-436.
[11]	邱春玲, 陶强, 范润龙, 王培智. 基于局部二进制描述SIFT特征的锆石图像配准方法[J]. 吉林大学学报(工学版), 2014, 44(6): 1793-1798.
[12]	顾播宇,孙俊喜,李洪祚,刘红喜,刘广文. 基于特征加权模块双方向二维主成分分析的人脸识别[J]. 吉林大学学报(工学版), 2014, 44(3): 828-833.
[13]	申铉京, 朱叶, 吕颖达, 陈海鹏. 基于SIFT和HSI模型的彩色图像复制-粘贴盲鉴别算法[J]. 吉林大学学报(工学版), 2014, 44(01): 171-176.
[14]	徐涛, 赵世佳, 张炜, 谭丽辉, 吕岗, 李恒. N重亏损系统摄动的原点移位组合近似方法[J]. 吉林大学学报(工学版), 2012, 42(增刊1): 147-150.
[15]	刘三民, 孙知信. 基于支持向量数据描述的P2P流量识别[J]. , 2012, 42(04): 947-951.

Metrics

Viewed

Full text

351

HTML			PDF

Just accepted	Online first	Issue	Just accepted	Online first	Issue
0	0	0	0	0	351

From	Others	local

Times	82	269
Rate	23%	77%

Abstract

283

Just accepted	Online first	Issue

0	0	283

	From	Others

	Times	283
	Rate	100%

Cited

Web of Science	Crossref	ScienceDirect	Search for Citations in Google Scholar >>


This page requires you have already subscribed to WoS.

Shared

Discussed