Elsevier

Neurocomputing

Volume 151, Part 3, 3 March 2015, Pages 1255-1261
Neurocomputing

MiLDA: A graph embedding approach to multi-view face recognition

https://doi.org/10.1016/j.neucom.2014.11.004Get rights and content

Abstract

In a vast number of real-world face recognition applications, gallery and probe image sets are captured from different scenarios. For such multi-view data, face recognition systems often perform poorly. To tackle this problem, in this paper we propose a graph embedding framework, which can project the multi-view data into a common subspace of higher discriminability between classes. This framework can be readily utilized to extend classical dimensionality reduction methods to multi-view scenarios. Hence, by utilizing the framework for multi-view face recognition, we propose multi-view linear discriminant analysis (MiLDA). We also empirically demonstrate that, for several distinct multi-view face recognition scenarios, MiLDA has an excellent performance and outperforms many popular approaches.

Section snippets

Canonical correlation analysis

The most famous subspace-based multi-view data analysis approach is canonical correlation analysis (CCA) [26], the goal of which is to extract pairs of directions that maximize correlation between two data sets.

Formally, given two matrices X1=[x1(1),x1(2),,x1(N)] in which x1(i)Rd1 and X2=[x2(1),x2(2),,x2(N)] in which x2(i)Rd2, CCA iteratively learns the pairs of normalized projection directions uncorrelated with the pairs previously learned. The first pair of projection directions v1 and v2

The proposed graph embedding framework for multi-view data analysis and MiLDA

In this section, we present our graph embedding framework for multi-view data analysis and MiLDA. First of all, we briefly review the graph embedding formulation, which unifies various classical dimensionality reduction approaches. We then introduce our graph embedding framework and elaborate its solution, and finally propose MiLDA on the basis of the framework.

Experiments

As we have mentioned, a big challenge in multi-view face recognition is due to the significant difference between face images captured in different scenarios. To empirically demonstrate the ability of MiLDA to overcome this challenge, here we conduct experiments for two multi-view face recognition scenarios, one is the multi-view face recognition across pose and illumination, and the other is the still-to-video face recognition.

Conclusions

In this paper, we have presented a graph embedding framework for multi-view data analysis. Our framework projects multi-view data into a common subspace of higher discriminability. By utilizing this framework, we have also proposed multi-view linear discriminant analysis (MiLDA) for multi-view face recognition. MiLDA can be solved efficiently, and has demonstrated superior empirical performance to some popular approaches. Future work will focus on its further improvement by combining it with

Acknowledgements

We thank the three anonymous reviewers for their constructive comments, in particular those on the relationship between MiLDA, MvDA and GMLDA. This work was partially supported by the National Basic Research Program of China 973 program under Grant no. 2013CB329403.

Yiwen Guo received the B.E. degree from Wuhan University, China, in 2011. He is currently working toward the Ph.D. degree with the Department of Electronic Engineering in Tsinghua University, China. His current research interests include computer vision, pattern recognition, machine learning and face recognition.

References (39)

  • X. Gao et al.

    Multi-view face recognition based on tensor subspace analysis and view manifold modeling

    Neurocomputing

    (2009)
  • Z. Zhao et al.

    Gabor face recognition by multi-channel classifier fusion of supervised kernel manifold learning

    Neurocomputing

    (2012)
  • A. Li, S. Shan, X. Chen, W. Gao, Maximizing intra-individual correlations for face recognition across pose differences,...
  • S. Li, X. Liu, X. Chai, H. Zhang, S. Lao, S. Shan, Morphable displacement field based image matching for face...
  • D. Yi, Z. Lei, S. Z. Li, Towards pose robust face recognition, in: IEEE Conference on Computer Vision and Pattern...
  • H.T. Ho et al.

    Pose-invariant face recognition using Markov random fields

    IEEE Trans. Image Process.

    (2013)
  • A. Asthana, T. Marks, M. Jones, K. Tieu, M. Rohith, Fully automatic pose-invariant face recognition via 3d pose...
  • S. Biswas et al.

    Pose-robust recognition of low-resolution face images

    IEEE Trans. Pattern Anal. Mach. Intell.

    (2013)
  • M. Kan, S. Shan, H. Zhang, S. Lao, X. Chen, Multi-view discriminant analysis, in: European Conference on Computer...
  • H. Han, S. Shan, X. Chen, S. Lao, W. Gao, Separability oriented preprocessing for illumination-insensitive face...
  • Z. Huang, X. Zhao, S. Shan, R. Wang, X. Chen, Coupling alignments with recognition for still-to-video face recognition,...
  • A. Sharma, D.W. Jacobs, Bypassing synthesis: PLS for face recognition with pose, low-resolution and sketch, in: IEEE...
  • A. Sharma, A. Kumar, H. Daume III, D.W. Jacobs, Generalized multiview analysis: a discriminative latent space, in: IEEE...
  • M. Kan et al.

    Domain adaptation for face recognition: targetize source domain bridged by common subspace

    Int. J. Comput. Vis.

    (2014)
  • X. Tang et al.

    Face sketch recognition

    IEEE Trans. Circuits Syst. Video Technol.

    (2004)
  • Q. Liu, X. Tang, H. Jin, H. Lu, S. Ma, Nonlinear approach for face sketch synthesis and recognition, in: IEEE...
  • X. Wang et al.

    Face photo-sketch synthesis and recognition

    IEEE Trans. Pattern Anal. Mach. Intell.

    (2009)
  • FaceVACS Software Developer Kit, Cognitec Systems GmbH, URL 〈http://www.cognitec-systems.de〉,...
  • W. Zhang, X. Wang, X. Tang, Coupled information-theoretic encoding for face photo-sketch recognition, in: IEEE...
  • Cited by (0)

    Yiwen Guo received the B.E. degree from Wuhan University, China, in 2011. He is currently working toward the Ph.D. degree with the Department of Electronic Engineering in Tsinghua University, China. His current research interests include computer vision, pattern recognition, machine learning and face recognition.

    Xiaoqing Ding received the B.E. degree from Tsinghua University, China, in 1962. She is currently a Professor and a Ph.D. Supervisor with the Department of Electronic Engineering, Tsinghua University. Her research interests include computer vision, pattern recognition, machine learning, image processing, face recognition, character recognition, etc. She was a recipient of a series of achievements on Chinese/multilanguage character recognition, face recognition, etc. She was a recipient of the most prestigious National Scientific and Technical Progress Awards in China in 1992, 1998, 2003, and 2008.

    Jing-Hao Xue received the B.Eng. degree in telecommunication and information systems in 1993 and the Dr.Eng. degree in signal and information processing in 1998, both from Tsinghua University, the M.Sc. degree in medical imaging and the M.Sc. degree in statistics, both from Katholieke Universiteit Leuven in 2004, and the degree of Ph.D. in statistics from the University of Glasgow in 2008. He has worked in the Department of Statistical Science at University College London as a Lecturer since 2008. His research interests include statistical and machine-learning techniques for pattern recognition, data mining and image processing, in particular supervised, unsupervised and incompletely-supervised learning for complex and high-dimensional data.

    View full text