[1]张毅,廖巧珍,罗元.融合二阶HOG与CS-LBP的头部姿态估计[J].智能系统学报编辑部,2015,10(5):741-746.[doi:10.11992/tis.201506019]
 ZHANG Yi,LIAO Qiaozhen,LUO Yuan.Head pose estimation fusing the second order HOG and CS-LBP[J].CAAI Transactions on Intelligent Systems,2015,10(5):741-746.[doi:10.11992/tis.201506019]
点击复制

融合二阶HOG与CS-LBP的头部姿态估计(/HTML)
分享到:

《智能系统学报》编辑部[ISSN:1673-4785/CN:23-1538/TP]

卷:
第10卷
期数:
2015年5期
页码:
741-746
栏目:
出版日期:
2015-10-25

文章信息/Info

Title:
Head pose estimation fusing the second order HOG and CS-LBP
作者:
张毅1 廖巧珍1 罗元2
1. 重庆邮电大学 自动化学院, 重庆 400065;
2. 重庆邮电大学 光电工程学院, 重庆 400065
Author(s):
ZHANG Yi1 LIAO Qiaozhen1 LUO Yuan2
1. College of Automation, Chongqing University of Posts and Telecommunications, Chongqing 400065, China;
2. College of Photoelectric Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China
关键词:
头部姿态估计梯度方向直方图(HOG)中心对称局部二值模式(CS-LBP)核主成分分析(KPCA)支持向量机(SVM)
Keywords:
head pose estimationhistogram of the orientation gradient (HOG)center symmetric local binary pattern (CS-LBP)kernel principal component analysis (KPCA)support vector machine (SVM)
分类号:
TP391.4
DOI:
10.11992/tis.201506019
文献标志码:
A
摘要:
针对头部姿态估计受光照变化、表情、噪声干扰等因素影响导致识别率低的问题,提出一种融合二阶梯度方向直方图(HOG)和中心对称局部二值模式(CS-LBP)特征的姿态特征,用于单帧图像的头部姿态估计。采用二阶HOG对人脸图像进行形状信息提取,得到人脸的轮廓特征;用CS-LBP进行局部纹理信息的提取,通过将二阶HOG提取的轮廓特征和CS-LBP提取的纹理特征进行融合,得到更有效的人脸特征;将融合的姿态特征通过核主成分分析(KPCA)变换非线性映射到高维核空间中,抽取其主元特征分量,采用支持向量机(SVM)分类器进行姿态估计。实验结果表明,方法和HOG、LBP、二阶HOG、CS-LBP方法相比有更高的分类准确率,对光照的变化有很好的鲁棒性。
Abstract:
In order to improve head pose recognition rate under variable illumination, expression, and noise, etc., a novel pose feature, fusing the second order histogram of the orientation gradient (HOG) with the center symmetric local binary pattern (CS-LBP) feature, is proposed in order to estimate head pose in a single frame image. The contour information of the facial image is extracted by the second order HOG, deriving the facial contour feature. CS-LBP is used to extract local texture information. More effective facial features can be obtained by fusing contour feature extracted by the second order HOG and the texture feature extracted by CS-LBP. Kernel principal component analysis (KPCA) is used to nonlinearly project the fused pose feature into a higher dimensional kernel space so as to further select the primary feature. A support vector machine (SVM) classifier is used for pose estimation. Experiment results show that the proposed method is more accurate than the HOG method and the LBP method. This method has good robustness for variable illumination.

参考文献/References:

[1] PATERAKI M, BALTZAKIS H, TRAHANIAS P. Visual estimation of pointed targets for robot guidance via fusion of face pose and hand orientation[C]//IEEE International Conference on Computer Vision Workshops. Barcelona, Spain, 2011:1060-1067.
[2] 李春玲, 邹北骥, 王磊. 基于面部和动作表情的双模态情绪强度估计[J]. 系统仿真学报, 2009, 21(16):5047-5052. LI Chunling, ZOU Beiji, WANG Lei. Double-mode estimation of emotion intensity based on facial and action’s expression[J]. Journal of System Simulation, 2009, 2l(16):5047-5052.
[3] MA B P, CHAI X J, WANG T J. A novel feature descriptor based on biologically inspired feature for head pose estimation[J]. Neurocomputing, 2013, 115:1-10.
[4] ZHANG Z Q, HU Y X, LIU M, et al. Head pose estimation in seminar room using multi view face detectors[M]. Heidelberg:Springer, 2007:299-304.
[5] 王毅, 叶德谦. 基于Gabor小波变换和两次DCT的人脸表情识别[J]. 微电子学与计算机, 2009, 26(5):262-264. WANG Yi, YE Deqian. Facial expression recognition based on Gabor and two times DCT[J]. Microelectronics & Computer, 2009, 26(5):262-264.
[6] JAIN A K, VAILAY A. Image retrieval using color and shape[J]. Pattern Recognition, 1996, 29(8):1233-1244.
[7] DALAL N, TRIGGS B. Histograms of oriented gradients for human detection[C]//IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR05). San Diego, USA, 2005:886-893.
[8] LU J W, PLATANIOTIS K N, VENETSANOPOULOS A N. Face recognition using kernel direct discriminant analysis algorithms[J]. IEEE Transactions on Neural Networks, 2003, 14(1):117-126.
[9] CAO H, YAMAGUCHI K, NAITO T, et al. Pedestrian recognition using second-order HOG feature[C]//Proceedings of 9th Asian Conference on Computer Vision (ACCV 2009). Xi’an, China, 2010:628-634.
[10] 董力赓, 陶霖密, 徐光祐. 基于二阶梯度朝向直方图特征的头部姿态估计[J]. 清华大学学报:自然科学版, 2011, 51(1):73-79. DONG Ligeng, TAO Linmi, XU Guangyou. Head pose estimation based on a second order histogram of the orientation gradient[J]. Journal of Tsinghua University:Science and Technology, 2011, 51(1):73-79.
[11] OJALA T, PIETIKANEN M, MAENPAA T. Multiresolution gray-scale and rotation invariant texture classification with local binary patterns[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002, 24(7):971-987.
[12] HEIKKILÄ M, PIETIKÄINEN M, SCHMID C. Description of interest regions with local binary patterns[J]. Pattern Recognition, 2009, 42(3):425-436.
[13] 张毅, 刘娇, 罗元, 等. 基于唇形的智能轮椅人机交互[J]. 控制工程, 2013, 20(3):501-505. ZHANG Yi, LIU Jiao, LUO Yuan, et al. Human-machine interaction based on shape of lip for intelligent wheelchair[J]. Control Engineering of China, 2013, 20(3):501-505.

备注/Memo

备注/Memo:
收稿日期:2015-06-11;改回日期:。
基金项目:国家自然科学基金资助项目(60905066).
作者简介:张毅, 男,1970年生,教授,博士生导师,主要研究方向为智能系统与移动机器人、机器人自主导航、机器视觉与模式识别、多传感器信息融合。主持并完成省部级及其他科研项目10余项,申请国家发明专利4项。发表论文60余篇,其中被SCI、EI、ISTP收录30余篇,出版专著1部,教材2部;廖巧珍,女,1989年生,硕士研究生,主要研究方向为模式识别和人机交互;罗元,女,1972年生,博士,教授,主要研究方向为机器人视觉导航、图像处理与模式识别。主持国家自然科学基金、重庆市自然科学基金、重庆市LED重大专项等国家和省部级科研项目10余项,获重庆市科技进步三等奖1项,申请和获得国家发明专利20余项。发表学术论文60余篇,其中被SCI、EI检索30余篇,出版专著4部。
通讯作者:廖巧珍.E-mail:490957008@qq.com.
更新日期/Last Update: 2015-11-16