JZUS - Journal of Zhejiang University SCIENCE

ENGINEERING Information Technology & Electronic Engineering

Accepted manuscript available online (unedited version)

Unseen head pose prediction using dense multivariate label distribution

Author(s): Gao-li Sang, Hu Chen, Ge Huang, Qi-jun Zhao
Affiliation(s): State Key Laboratory of Fundamental Science on Synthetic Vision, College of Computer Science, Sichuan University, Chengdu 610064, China; more
Corresponding email(s): g.sang@foxmail.com, huchen@scu.edu.cn, 26434368@qq.com, qjzhao@scu.edu.cn
Key Words: Head pose estimation, Dense multivariate label distribution, Sampling intervals, Inconsistent labels

Share this article to： More <<< Previous Paper \|Next Paper >>>

Gao-li Sang, Hu Chen, Ge Huang, Qi-jun Zhao. Unseen head pose prediction using dense multivariate label distribution[J]. Frontiers of Information Technology & Electronic Engineering,in press.https://doi.org/10.1631/FITEE.1500235

@article{title="Unseen head pose prediction using dense multivariate label distribution",
author="Gao-li Sang, Hu Chen, Ge Huang, Qi-jun Zhao",
journal="Frontiers of Information Technology & Electronic Engineering",
year="in press",
publisher="Zhejiang University Press & Springer",
doi="https://doi.org/10.1631/FITEE.1500235"
}

%0 Journal Article
%T Unseen head pose prediction using dense multivariate label distribution
%A Gao-li Sang
%A Hu Chen
%A Ge Huang
%A Qi-jun Zhao
%J Frontiers of Information Technology & Electronic Engineering
%P 516-526
%@ 2095-9184
%D in press
%I Zhejiang University Press & Springer
doi="https://doi.org/10.1631/FITEE.1500235"

TY - JOUR
T1 - Unseen head pose prediction using dense multivariate label distribution
A1 - Gao-li Sang
A1 - Hu Chen
A1 - Ge Huang
A1 - Qi-jun Zhao
J0 - Frontiers of Information Technology & Electronic Engineering
SP - 516
EP - 526
%@ 2095-9184
Y1 - in press
PB - Zhejiang University Press & Springer
ER -
doi="https://doi.org/10.1631/FITEE.1500235"

Abstract
Chinese Summary
Academic Network
Reviewer Comment

Abstract: Accurate head poses are useful for many face-related tasks such as face recognition, gaze estimation, and emotion analysis. Most existing methods estimate head poses that are included in the training data (i.e., previously seen head poses). To predict head poses that are not seen in the training data, some regression-based methods have been proposed. However, they focus on estimating continuous head pose angles, and thus do not systematically evaluate the performance on predicting unseen head poses. In this paper, we use a dense multivariate label distribution (MLD) to represent the pose angle of a face image. By incorporating both seen and unseen pose angles into MLD, the head pose predictor can estimate unseen head poses with an accuracy comparable to that of estimating seen head poses. On the Pointing’04 database, the mean absolute errors of results for yaw and pitch are 4.01° and 2.13°, respectively. In addition, experiments on the CAS-PEAL and CMU Multi-PIE databases show that the proposed dense MLD-based head pose estimation method can obtain the state-of-the-art performance when compared to some existing methods.

Abstract: This paper proposes a head pose estimation method using dense multivariate label distribution. It solves the problem that the training data cannot cover all the possible test data due to large (head pose) sampling interval in training. The key idea is to produce a dense MLD to sample head pose angles densely. The results appear quite promising.

基于稠密多变量标签的“连续”头部姿态估计方法

目的：精确的头部姿态估计对于人脸相关的应用，如人脸识别、视线估计、情感分析等具有重要意义。大多数现有的人脸姿态估计方法仅能对训练数据库包含姿态的情况进行估计。为实现对训练数据库不包含姿态的情况进行预测，有学者提出了基于回归的头部姿态估计方法。然而，这些基于回归的方法虽然可以预测连续的姿态，但是却很少有相关的系统性性能评估。
方法：针对训练数据库不包含姿态的估计问题，本文提出使用稠密多变量标签分布表示人脸姿态。通过给样本分配稠密化的多变量标签，可以实现对数据库不包含姿态的情况进行较为准确的估计。
结论：本文方法在Pointing’04数据库上的yaw和pitch方向分别取得了平均绝对误差4.01°和2.13°。此外，在CAL-PEAL，Multi-PIE等公开库上的实验表明，本文方法在训练数据库包含姿态上的预测性能也优于其他比较先进的方法。

关键词组：头部姿态估计；稠密多变量标签分布；角度间隔；不一致性标签

Darkslateblue:Affiliate; Royal Blue:Author; Turquoise:Article

Reference

[1]Aghajanian, J., Prince, S.J.D., 2009. Face pose estimation in uncontrolled environments. Proc. British Machine Vision Conf., p.1-11.

[2]Berger, A.L., Pietra, V.J.D., Pietra, S.A.D., 1996. A maximum entropy approach to natural language processing. Comput. Ling., 22(1):39-71.

[3]Bowyer, K.W., Chang, K., Flynn, P., 2006. A survey of approaches and challenges in 3D and multi-modal 3D+2D face recognition. Comput. Vis. Image Understand., 101(1):1-15.

[4]Brunelli, R., 1997. Estimation of pose and illuminant direction for face processing. Image Vis. Comput., 15(10):741-748.

[5]Cai, Y., Yang, M.L., Li, Z.Q., 2015. Robust head pose estimation using a 3D morphable model. Math. Prob. Eng., 2015:678973.1-678973.10.

[6]Do, M.N., 2003. Fast approximation of Kullback-Leibler distance for dependence trees and hidden Markov models. IEEE Signal Process. Lett., 10(4):115-118.

[7]Fenzi, M., Leal-Taixé, L., Rosenhahn, B., et al., 2013. Class generative models based on feature regression for pose estimation of object categories. Proc. IEEE Conf. on Computer Vision and Pattern Recognition, p.755-762.

[8]Fitzpatrick, P., 2000. Head Pose Estimation Without Manual Initialization. Report, Massachusetts Institute of Technology, Cambridge.

[9]Gao, W., Cao, B., Shan, S.G., et al., 2008. The CAS-PEAL large-scale Chinese face database and baseline evaluations. IEEE Trans. Syst. Man Cybern. A, 38(1):149-161.

[10]Geng, X., Xia, Y., 2014. Head pose estimation based on multivariate label distribution. Proc. IEEE Conf. on Computer Vision and Pattern Recognition, p.1837-1842.

[11]Gourier, N., Hall, D., Crowley, J.L., 2004. Estimating face orientation from robust detection of salient facial features. Proc. Int. Workshop on Visual Observation of Deictic Gestures. Available from http://www-prima.inrialpes.fr/perso/Gourier/Faces/HPDatabase.html.

[12]Gross, R., Matthews, I., Cohn, J., et al., 2010. Multi-PIE. Image Vis. Comput., 28(5):807-813.

[13]Haj, M.A., Gonzàlez, J., Davis, L.S., 2012. On partial least squares in head pose estimation: how to simultaneously deal with misalignment. Proc. IEEE Conf. on Computer Vision and Pattern Recognition, p.2602-2609.

[14]Hu, C.L., Gong, L.Y., Wang, T.J., et al., 2014. An effective head pose estimation approach using Lie algebrized Gaussians based face representation. Multim. Tools Appl., 73(3):1863-1884.

[15]Huang, D., Storer, M., de la Torre, F., et al., 2011. Supervised local subspace learning for continuous head pose estimation. Proc. IEEE Conf. on Computer Vision and Pattern Recognition, p.2921-2928.

[16]Jain, V., Crowley, J.L., 2013. Head pose estimation using multi-scale Gaussian derivatives. Proc. 18th Scandinavian Conf. on Image Analysis, p.319-328.

[17]Krüger, V., Sommer, G., 2002. Gabor wavelet networks for efficient head pose estimation. Image Vis. Comput., 20(9-10):665-672.

[18]Liu, D.C., Nocedal, J., 1989. On the limited memory BFGS method for large scale optimization. Math. Program., 45(1):503-528.

[19]Lu, F., Sugano, Y., Okabe, T., et al., 2012. Head pose-free appearance-based gaze sensing via eye image synthesis. Proc. 21st Int. Conf. on Pattern Recognition, p.1008-1011.

[20]Lu, F., Okabe, T., Sugano, Y., et al., 2014. Learning gaze biases with head motion for head pose-free gaze estimation. Image Vis. Comput., 32(3):169-179.

[21]Ma, B.P., Chai, X.J., Wang, T.J., 2013. A novel feature descriptor based on biologically inspired feature for head pose estimation. Neurocomputing, 115:1-10.

[22]Ma, B.P., Li, A.N., Chai, X.J., et al., 2014. CovGa: a novel descriptor based on symmetry of regions for head pose estimation. Neurocomputing, 143:97-108.

[23]Ma, B.P., Huang, R., Qin, L., 2015. VoD: a novel image representation for head yaw estimation. Neurocomputing, 148:455-466.

[24]Ma, X.H., Tan, Y.Q., Zheng, G.M., 2013. A fast classification scheme and its application to face recognition. J. Zhejiang Univ.-Sci. C (Comput. & Electron.), 14(7):561-572.

[25]Murphy-Chutorian, E., Trivedi, M.M., 2009. Head pose estimation in computer vision: a survey. IEEE Trans. Patt. Anal. Mach. Intell., 31(4):607-626.

[26]Pang, H., Lin, A., Holford, M., et al., 2006. Pathway analysis using random forests classification and regression. Bioinformatics, 22(16):2028-2036.

[27]Sim, T., Baker, S., Bsat, M., 2002. The CMU pose, illumination, and expression (PIE) database. Proc. 5th IEEE Int. Conf. on Automatic Face and Gesture Recognition, p.46-51.

[28]Tang, Y.Q., Sun, Z.N., Tan, T.N., 2014. A survey on head pose estimation. Patt. Recogn. Artif. Intell., 27(3):213-225.

[29]Wu, J.W., Trivedi, M.M., 2008. A two-stage head pose estimation framework and evaluation. Patt. Recog., 41(3):1138-1158.

[30]Zhang, Z.P., Luo, P., Loy, C.C., et al., 2014. Facial landmark detection by deep multi-task learning. Proc. 13th European Conf. on Computer Vision, p.94-108.

[31]Zhu, R.H., Sang, G.L., Cai, Y., et al., 2013. Head pose estimation with improved random regression forests. Proc. 8th Chinese Conf. on Biometric Recognition, p.457-465.

[32]Zhu, X.X., Ramanan, D., 2012. Face detection, pose estimation, and landmark localization in the wild. Proc. IEEE Conf. on Computer Vision and Pattern Recognition, p.2879-2886.

Open peer comments: Debate/Discuss/Question/Opinion

<1>

- Go to

基于稠密多变量标签的“连续”头部姿态估计方法

Darkslateblue:Affiliate; Royal Blue:Author; Turquoise:Article

Reference