ZTE

ZTE Communications ›› 2014, Vol. 12 ›› Issue (4): 23-29.doi: DOI:10.3969/j.issn.1673-5188.2014.04.004

• Special Topic • Previous Articles     Next Articles

Facial Landmark Localization by Gibbs Sampling

Bofei Wang1, Diankai Zhang1, Chi Zhang2, Jiani Hu2, and Weihong Deng2   

  1. 1. ZTE Corporation, Shenzhen 518057, China;
    2. Beijing University of Posts and Telecommunication, Beijing 100876, China
  • Received:2014-08-20 Online:2014-12-25 Published:2014-12-25
  • About author:Bofei Wang (wang.bofei@zte.com.cn) received his BE degree in electronic information engineering and MS in Communication and information system from Huazhong University of Science and Technology (HUST), China in 2003 and 2007. He is a senior video and image algorithm engineer of ZTE Corporation. His research interests include video and image processing, pattern recognition, and computer vision.

    Diankai Zhang (zhang.diankai@zte.com.cn) received his BE degree in electronic information engineering and MS degree in signal and information processing from Nanjing University of Posts and Telecommunications (NUPT), China in 2006 and 2009. He is a senior video and image algorithm engineer of ZTE Corporation. His research interests include video and image processing, pattern recognition, and computer vision.

    Chi Zhang (zhangchi2013@bupt.edu.cn) received his BE degree in Electronic Information Engineering from NUPT in 2013, and is currently a master student in School of Information and Telecommunications Engineering of Beijing University of Posts and Telecommunications (BUPT), China. His research interests include pattern recognition, machine learning, and computer vision.

    Jiani Hu (jnhu@bupt.edu.cn) received her BE degree in telecommunication engineering from China University of Geosciences in 2003, and PhD degree in signal and information processing from BUPT in 2008. She is currently a lecturer in School of Information and Telecommunications Engineering, BUPT. Her research interests include information retrieval, statistical pattern recognition, and computer vision.

    Weihong Deng (whdeng@bupt.edu.cn) is an associate professor in School of Information and Telecommunications Engineering, BUPT. His research interests include statistical pattern recognition and computer vision, with a particular emphasis on face recognition. He has published over 40 papers in international journals and conferences, including a technical comment on face recognition in Science magazine. He also serves as the reviewer for several international journals, such as IEEE TPAMI, IJCV, IEEE TIP, IEEE TIFS, PR, and IEEE TSMC - B. His dissertation titled “Highly accurate face recognition algorithms”was awarded the Outstanding Doctoral Dissertation by Beijing Municipal Commission of Education in 2011. He has been supported by the program for New Century Excellent Talents by the Ministry of Education of China since 2013.
  • Supported by:
    This work is supported by ZTE Industry-Academia-Research Cooperation Funds.

PDF (PC)

Abstract: In this paper, we introduce a novel method for facial landmark detection. We localize facial landmarks according to the MAP criterion. Conventional gradient ascent algorithms get stuck at the local optimal solution. Gibbs sampling is a kind of Markov Chain Monte Carlo (MCMC) algorithm. We choose it for optimization because it is easy to implement and it guarantees global convergence. The posterior distribution is obtained by learning prior distribution and likelihood function. Prior distribution is assumed Gaussian. We use Principle Component Analysis (PCA) to reduce the dimensionality and learn the prior distribution. Local Linear Support Vector Machine (LL-SVM) is used to get the likelihood function of every key point. In our experiment, we compare our detector with some other well-known methods. The results show that the proposed method is very simple and efficient. It can avoid trapping in local optimal solution.

Key words: facial landmarks, MAP, Gibbs sampling, MCMC, LL-SVM