Wei-Shi Zheng 

Associate Professor

 

Key Member of Guangdong Province Introduced Innovative Computing Science Team

(广东省引进创新科研团队计算科学科研团队核心成员)

 

School of Information Science and Technology 

Sun Yat-sen University, China

E-mail: wszheng@ieee.org

Web: http://sist.sysu.edu.cn/~zhwshi/

Google Scholar

Previous: Department of Computer Science, Queen Mary University of London, United Kingdom




All of you who really like to do research on artificial intelligence are welcomed! It does not matter whether you are a graduate student or just an undergraduate; as long as you are willing to work hard, you are welcomed to join my research team and I will enjoy immensely working with you!

[
Lecture Notes (Sun Yat-sen University)][@微博][My Photos]
[
For Prospective Graduate Students(研究生招生)]
[Research for Undergraduate Students(本科生科研计划)]
[Referene Letters(推荐信)][Undergraduate Thesis(本科毕业论文)]


 

Research Direction: Machine Vision and Intelligence Learning (机器视觉与智能学习)

Recent Research Focus

1. Machine Learning (机器学习):
    (1) Large Scale Learning (大规模学习);
    (2) Transfer Learning(迁移学习);
    (3) Metric Learning (距离学习)
.

2. Image Understanding (图像理解):
    (1) Face Recognition (人脸识别);
    (2) Person Re-identification (行人再标识);
    (3) Action/Activity Recognition (动作及行为识别)


Brief Bio
My English name is Jason. I have joint SUN YAT-SEN University since Jan. 2011 under the one-hundred-people program. My early research was on face recognition based on subpsace methods, when I worked for Ph.D. degree in Applied Mathematics with Prof. Jianhuang Lai and Prof. Pong C. Yuen at Sun Yat-sen University and Hong Kong Baptist University. I have been a visiting student working with Prof. Stan Z. Li at the Institute of Automation, Chinese Academy of Sciences. There, I started my research on sparsity learning for pattern recognition with sparse one-sided non-negativity matrix factorisation. After 2008, I worked as a Postdoctoral Researcher on the European SAMURAI project for person association with Prof. Shaogang Gong  and Dr. Tao Xiang. In Queen Mary University of London, I proposed to use relative comparison to overcome the ill-posed matching problem between the appearance of people across non-overlapping camera views. The idea was also extended to group of people association. I also created an i-LIDS dataset for evaluating this matching problem.

After I joined Sun Yat-sen University in 2011, I become focusing on large scale machine learning research. One motivation for me to consider this research is because the relative comparison model I proposed is
still not scalable to very large scale data. In the big picture, data become more and more available nowadays. The so-called Big Data becomes an important research issue now. So, I aim at developing large scale machine learning model in order to make algorithm/model scalable to Big Data. Meanwhile, I hope my current machine learning research would help solve the vision problems, such as human activity. As a team member of Key Member of Guangdong Province Introduced Innovative Computing Science Team, we have full support on this research.



Important  Honours

- Guangdong Natural Science Funds for Distinguished Young Scholars (广东省自然科学杰出青年基金)
- New Star of science and technology of Guangzhou (广州市珠江科技新星)
- One-hundred People Program of Sun Yat-sen University (中山大学百人计划)



Teaching

- Advanced Algebra (in both Chinese and English, first semester)
- Linear Algebra (2+2 international undergraduate program, all in English, first semester)
- Engineering Mathematics (2+2 international undergraduate program, all in English, second semester)
- Mathematics Comprehensive Training (in Chinese, third semester(for two weeks))


Tutorials

1. ICCV 2013 tutorial: Sparsity Estimation and Robust Learning: A Half-quadratic Minimization View. (With Dr. Ran He, Prof. Liang Wang)
2. ICPR 2012 tutorial: Half-quadratic Optimization for Sparsity Estimation and Robust Learning. (With Dr. Ran He, Prof. Liang Wang) [Download PDF]
3. ACCV 2012 tutorial: Half-quadratic Optimization for Computer Vision. (With Dr. Ran He, Prof. Liang Wang) [Download PDF]
4. CCBR 2013 tutorial: Person Re-identification.[Download PDF]


News: 1 TIFS paper

Selected Publications (Full Publication List) (Grouped by Topic)

(*: supervisor of the students of Sun Yat-sen University)

Book & Chapter

 

1. Wei-Shi Zheng, Zhenan Sun, Yunhong Wang, Xilin Chen, Pong C. Yuen, and Jianhuang Lai, "Biometric Recognition," LNCS 7701, Springer, 2012 (ISBN: 978-3-642-35135-8)

2. Wei-Shi Zheng, JianHuang Lai, Pong C. Yuen, "Linear Dimension Reduction," Encyclopedia of Biometrics, Springer, pp. 899-903, 2009.

3. Wei-Shi Zheng, JianHuang Lai, Xiaohua Xie, Yan Liang, Pong C. Yuen, and Yaoxian Zou, "Kernel Methods for Facial Image Preprocessing," Pattern Recognition, Machine Intelligence and Biometrics---Expanding Frontiers, Springer and High Education Press, 2011.

4. Wei-Shi Zheng, Shaogang Gong, Tao Xiang, “Group Association: Assisting Re-identification by Visual Context,” Person Re-Identification, Springer, 2013.


 

Journal Publications:

 

17. Wei-Shi Zheng, Shaogang Gong, and Tao Xiang, "Re-identification by Relative Distance Comparison," IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 35, no. 3, pp. 653-668, 2013. [Code][meta data for ilids, ethz and viper used in the PAMI paper]

16.Ran He, Wei-Shi Zheng, Tieniu Tan, and Zhenan Sun, "Half-quadratic based Iterative Minimization for Robust Sparse Representation," IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 36, no. 2, pp. 261-275, 2014.

15. Wei-Shi Zheng, Shaogang Gong, and Tao Xiang, "Quantifying and Transferring Contextual Information in Object Detection," IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 34, no. 4, pp. 762-777, 2012.

14. Ran He, Wei-Shi Zheng, and BaoGang Hu, "Maximum Correntropy Criterion for Robust Face Recognition," IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 33, no. 8, pp. 1561 - 1576, 2011.

13. Jun-Yong Zhu (student), Wei-Shi Zheng*, Jian-Huang Lai, Stan Z. Li, "Matching NIR Face to VIS Face using Transduction," IEEE Transactions on Information Forensics and Security, vol. 9, no. 3, pp. 501-514, 2014.

12. Wei-Shi Zheng, JianHuang Lai, Shengcai Liao, and Ran He, "Extracting Non-negative Basis Images Using Pixel Dispersion Penalty," Pattern Recognition, vol. 45, no. 8, pp. 2912-2926, 2012.[CODE]

11. Ran He, Wei-Shi Zheng, Bao-Gang Hu, and Xiang-Wei Kong, "Two-stage Nonnegative Sparse Representation for Large-scale Face Recognition," IEEE Trans. on Neural Networks and Learning System (TNNLS), accepted, 2012.

10.Yu Chen+, Wei-Shi Zheng+, Jian-Huang Lai, "Discriminant Subspace Learning Constrained by Locally Statistical Uncorrelation for Face Recognition," Neural Networks, to appear, 2013.

9. Wei-Shi Zheng, JianHuang Lai, and Pong C. Yuen, "Penalized Pre-image Learning in Kernel Principal Component Analysis," IEEE Trans. on Neural Networks, vol. 21, no. 4, pp. 551-570, 2010. []

8. Wei-Shi Zheng, JianHuang Lai, Pong C. Yuen, and Stan Z. Li, "Perturbation LDA: Learning the Difference between the Class Empirical Mean and Its Expectation," Pattern Recognition, vol. 42, no. 5, pp. 764-779, 2009. []

7. Wei-Shi Zheng, JianHuang Lai, and Stan Z. Li, "1D-LDA versus 2D-LDA: When Is Vector-based Linear Discriminant Analysis Better than Matrix-based?" Pattern Recognition, vol. 41, no. 7, pp. 2156-2172, 2008.

6Wei-Shi Zheng, JianHuang Lai, and Pong C. Yuen, "GA-fisher: A New LDA-based Face Recognition Algorithm with Selection of Principal Components," IEEE Transactions on Systems, Man, and Cybernetics, Part B, vol. 35, no. 5, pp. 1065-1078, 2005.

5. Xiaohua Xie, Wei-Shi Zheng, JianHuang Lai, Pong C. Yuen, and Ching Y. Suen, "Normalization of Face Illumination Based on Large- and Small- Scale Features," IEEE Trans. on Image Processing, vol. 20, no. 7, pp. 1807 - 1821, 2011.

4. Ran He, Wei-Shi Zheng, and BaoGang Hu, and Xiang-Wei Kong, "A Regularized Correntropy Framework for Robust Pattern Recognition," Neural Computation, vol. 23, no. 8, pp. 2074-2100, 2011.

3. Guoyun Lian, JianHuang Lai, and Wei-Shi Zheng, "Spatial-temporal Consistent Labeling of Tracked Pedestrians across Non-overlapping Camera Views," Pattern Recognition, vol. 44, no. 5, pp. 1121-1136, 2011.

2. Xiaohua Xie, JianHuang Lai, and Wei-Shi Zheng, "Extraction of Illumination Invariant Facial Features from A Single Image Using Nonsubsampled Contourlet Transform," Pattern Recognition, vol. 43, no.12, pp. 41774189, 2010.

1. Ran He, BaoGang Hu, Wei-Shi Zheng, and Xiang-Wei Kong, "Robust Principal Component Analysis Based on Maximum Correntropy Criterion," IEEE Trans. on Image Processing, vol. 20, no. 6, pp. 1485-1494, 2011.

 

more ...

 

Top Conferences:

 

15. Jian-Fang Hu (student), Wei-Shi Zheng*, Jian-huang Lai, Shaogang Gong, and Tao Xiang, "Recognising Human-Object Interaction via Exemplar based Modelling," IEEE Conf. on Computer Vision (ICCV) 2013, to appear

14. Yuanlu Xu, Liang Lin, Wei-Shi Zheng, and Xiaobai Liu, "Human Re-identification by Matching Compositional Template with Cluster Sampling,"  IEEE Conf. on Computer Vision (ICCV) 2013, to appear

13. Jian-sheng Wu (student), Wei-Shi Zheng*, Jian-huang Lai, "Euler Clustering," International Joint Conference on Artificial Intelligence (IJCAI), 2013.

12.Longkai Huang(undergraduate student), Qiang Yang(student), Wei-Shi Zheng*, "Online Hashing," International Joint Conference on Artificial Intelligence (IJCAI), 2013. [(including labelme dataset)]

11. Qiang Yang(student), Longkai Huang(undergraduate student), Wei-Shi Zheng*, Yingbiao Ling, "Smart Hashing Update for Fast Response," International Joint Conference on Artificial Intelligence (IJCAI), 2013.

10. Wenqi Li (co-supervised student), Jianguo Zhang, Stephen McKenna, Wei-Shi Zheng, Maria Coats, Frank Carey "Learning from Partially Annotated OPT Images by Contextual Relevance Ranking," International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), 2013, accepted. (This paper is for the NSFC-RSE joint project. Wenqi Li is the co-supervised student when I was in University of Dundee)

9. Wei-Shi Zheng, Shaogang Gong, and Tao Xiang, "Transfer Re-identification: From Person to Set-based Verification", IEEE Conf. on Computer Vision and Pattern Recognition, Rhode Island, USA, June 2012.

8. Wei-Shi Zheng, Shaogang Gong, and Tao Xiang, "Person Re-identification by Probabilistic Relative Distance Comparison", IEEE Conf. on Computer Vision and Pattern Recognition, 2011 (Acceptance rate: 26.4%)

7Wei-Shi Zheng, Shaogang Gong, and Tao Xiang, "Quantifying Contextual Information for Object Detection," In Proc. IEEE International Conference on Computer Vision (ICCV), Kyoto, Japan, October, 2009. (Acceptance rate: 23%)

6Wei-Shi Zheng, Stan Z. Li, JianHuang Lai, and Shengcai Liao, “On Constrained Sparse Matrix Factorization,” In Proc. IEEE International Conference on Computer Vision (ICCV), Brazil, 2007. (Acceptance rate: 23%)

5. Ran He and Wei-Shi Zheng, "Nonnegative Sparse Coding for Discriminative Semi-supervised Learning," IEEE Conf. on Computer Vision and Pattern Recognition, 2011. (Acceptance rate: 26.4%)

4. Xiaohua Xie, Wei-Shi Zheng, JianHuang Lai, and Pong C. Yuen, "Face Illumination Normalization on Large and Small Scale Features," In Proc. IEEE International Conference on Computer Vision and Pattern Recognition (CVPR), 2008. (Acceptance rate: 31%)

3. Ran He, BaoGang Hu, Wei-Shi Zheng, and YanQing Guo, "Two-stage Sparse Representation for Robust Recognition on Large-scale Database," In Proc. Twenty-Fourth AAAI Conference on Artificial Intelligence , 2010 (oral). (Acceptance rate: 26.9%)

2. Ran He, Tieniu Tan, Zhenan Sun, and Wei-Shi Zheng, "Recovery of Corrupted Low-Rank Matrices via Half-Quadratic based Nonconvex Minimization," IEEE Conf. on Computer Vision and Pattern Recognition, 2011. (Acceptance rate: 26.4%)

1. Ran He, Tieniu Tan, Liang Wang, and Wei-Shi Zheng, "l2,1 Regularized Correntropy for Robust Feature Selection", IEEE Conf. on Computer Vision and Pattern Recognition, Rhode Island, USA, June 2012.

 

Main Conferences in Computer Vision and Machine Learning:

 

1. Jun-Yong Zhu (student), Wei-Shi Zheng*, JianHuang Lai, " Logarithm Gradient Histogram: A General Illumination Invariant Descriptor for Face Recognition," IEEE Conference on Automatic Face and Gesture Recognition, 2013 (oral).[CODE]

2. Jun-Yong Zhu(student), Wei-Shi Zheng*, JianHuang Lai, "Transductive VIS-NIR Face Matching," IEEE International Conference on Image Processing, 2012.

3. Wei-Shi Zheng, Shaogang Gong, and Tao Xiang, "Unsupervised Selective Transfer Learning for Object Recognition," Asian Conference on Computer Vision (ACCV) 2010.

4Wei-Shi Zheng, Shaogang Gong, and Tao Xiang, "Associating Groups of People," In Proc. British Machine Vision Conference (BMVC), London, September 2009.

 

 

5Wei-Shi Zheng, Jianhuang Lai, "Regularized Locality Preserving Learning of Pre-Image Problem in Kernel Principal Component Analysis, " In Proc. 18th International Conference on Pattern Recognition (ICPR), 2006. (top 15%, oral)

 

 

6. Wei-Shi Zheng, JianHuang Lai, and Pong C. Yuen, “Weakly Supervised Learning on Pre-image Problem in Kernel Methods,” In Proc. 18th International Conference on Pattern Recognition (ICPR), 2006

 

more ...

 

Academic Activity

- Young Associate Editor, Frontiers of Computer Science
-
Area Chair, IEEE Conference Series on Advanced Video and Signal based Surveillance, Beijing, 2012.
- Programme Chair, Chinese Confrence on Biometric Recognition (CCBR), 2012.
-
Programme Committee Members: BTAS, ICPR, BMVC, ACCV
......

- Reviewer for:
  * IEEE Trans. on Pattern Analysis and Machine Intelligence.
  * IEEE Trans. on Neural Networks.
  * Pattern Recognition.
  * IEEE Trans. on Circuits and Systems for Video Technology.
......

 

Main Research (snapshot): Person/Object Association & Face Recognition

Human Body-based ASSOCIATION
 

 

Context Quantification for Object Detection

>> Context is critical for minimising ambiguity in object detection. In this work, a novel context modelling framework is proposed without the need of any prior scene segmentation or context annotation. This is achieved by  exploring a new polar geometric histogram descriptor for context representation. In order to quantify context, we formulate a new context risk function and a maximum margin context (MMC) model to solve the minimization problem of the risk function. Crucially, the usefulness and goodness of contextual information is evaluated directly and explicitly through a discriminant context inference method and a context confidence function, so that only reliable contextual information that is relevant to object detection is utilised.

---
W.-S. Zheng, S. Gong, and T. Xiang, "," ICCV 2009. []

Wei-Shi Zheng, Shaogang Gong, and Tao Xiang, "Quantifying and Transferring Contextual Information in Object Detection," accepted by IEEE Trans. on Pattern Analysis and Machine Intelligence, 2011.

 

  Group of People Matching

In a crowded public space, people often walk in groups, either with people they know or strangers. Associating a group of people over space and time can assist understanding individual's behaviours as it provides vital visual context for matching individuals within the group. Seemingly an `easier' task compared with person matching, this problem is in fact very challenging because a group of people can be highly non-rigid with changing relative position of people within the group and severe self-occlusions. For the first time, the problem of matching/associating groups of people over large space and time captured in multiple non-overlapping camera views is addressed by us. Specifically, a novel people group representation and a group matching algorithm are proposed. The former addresses changes in the relative positions of people in a group and the latter deals with variations in illumination and viewpoint across camera views. We also demonstrate a notable enhancement on individual Person matching by utilising the group description as visual context.

---
W.-S. Zheng, S. Gong, and T. Xiang, "Associating Groups of People," BMVC 2009. []

 

 

Person Re-identification

Developed a probabilistic relative distance comparison model for matching individual person across non-overlapping camera views in a surveillance system. The advantage of this model is that rather than maximizing the difference in distances between a pair of relevant person images and any related irrelevant pair, the proposed model is a soft discriminant model and only directly minimizes the probabilistic error of the relative distance comparison between them, which is proved to be more effective for person re-identification in order to alleviate the over-fitting problem due to the large variations of person’s appearance across non-overlapping camera views

---
W.-S. Zheng, S. Gong, and T. Xiang, "Person Re-identification by Probabilistic Relative Distance Comparison", IEEE Conf. on Compuer Vision and Pattern Recognition, 2011. [PDF] 

Wei-Shi Zheng, Shaogang Gong, and Tao Xiang, "Re-identification by Relative Distance Comparison," IEEE Trans. on Pattern Analysis and Machine Intelligence,  2013.[PDF]

[meta data for ilids, ethz and viper used in the PAMI paper][Code]

i-LIDS dataset

 

  Kernel Adaptation Transfer

A kernel adaptation transfer technique that utilizes unlabelled auxiliary data to quantify and select the most relevant transferrable knowledge for recognizing a target class object against the background given very limited target training samples. The advantages of this model are that it does not assume that the auxiliary data are labelled nor the relationships between target and auxiliary classes are known a priori, and it performs selective transfer of knowledge extracted from the auxiliary data to minimize the negative knowledge transfer suffered by many existing methods.

---
Wei-Shi Zheng, Shaogang Gong, and Tao Xiang, "Unsupervised Selective Transfer Learning for Object Recognition," Asian Conference on Computer Vision (ACCV) 2010.[PDF]

 

FACE RECOGNITION
         1. Linear Discriminant Feature Extraction (Subspace Methods))
  Principal Component Selection

>> There is some argument for principal component selection in PCA+LDA. This work shows small principal components (corresponding to small eigenvalues) are useful and should be carefully selected in PCA+LDA. A foundation of principal component selection in LDA is established. New GA technique is used for implementation.

---
Wei-Shi Zheng, J. H. Lai, and Pong C. Yuen, “,” IEEE Transactions on Systems, Man, and Cybernetics, Part B, vol. 35, no. 5, pp. 1065-1078, 2005. []
 
1D-LDA vs. 2D-LDA

>> From 2003 to 2008, lots of work have shown that algorithms with (2D) matrix-based representation perform better than the traditional (1D) vector-based ones. Specially, 2D-LDA was widely reported to outperform 1D-LDA. However, would the matrix-based linear discriminant analysis be always superior and when would 1D-LDA be better? This work gives some impressive theoretical analysis and experimental comparison between 1D-LDA and 2D-LDA. Different from existing views, we find that there is no convinced evidence that 2D-LDA would always outperform 1D-LDA when the number of training samples for each class is small or when the number of discriminant features used is small.

---
Wei-Shi Zheng, J. H. Lai, and Stan Z. Li, “” Pattern Recognition, vol. 41, no. 7, pp. 2156-2172, 2008. []
  Perturbation LDA

>> In deriving the Fisher’s LDA formulation, there is an assumption that the class empirical mean is equal to its expectation. However, this may not be valid in practice and this problem has been rarely discussed before. From the "perturbation" perspective, we develop a new algorithm, called perturbation LDA (P-LDA), in which perturbation random vectors are introduced to learn the effect of the difference between the class empirical mean and its expectation in Fisher criterion.

---
Wei-Shi Zheng, J. H. Lai, Pong C. Yuen, and Stan Z. Li, "," Patten Recognition, vol. 42, no. 5,  pp. 764-779, 2009. [] []

2. Sparse Feature Learning for Classification

 


Sparse One-sided Non-negative Matrix Factorization

>> NMF, which is a two-sided non-negativity based matrix factorization, is popular for extraction of sparse features. However, why non-negativity should be imposed on both components and coefficients? What is case if some constraint is released? In this work, we find releasing the non-negativity constraint on the coefficient term in NMF would help extract equally/much sparser and more reconstrutive components/features as compared to the two-sided non-negativity matrix factorization techniques.
The exact 17 local components of Swimmer data set are successfully extracted for the first time (to our best knowledge).

---
Wei-Shi Zheng, Stan Z. Li, J. H. Lai, and Shengcai Liao, “,” 11th IEEE International Conference on Computer Vision (ICCV), 2007. []

Wei-Shi Zheng, JianHuang Lai, Shengcai Liao, and Ran He, "Extracting Non-negative Basis Images Using Pixel Dispersion Penalty," Pattern Recognition, vol. 45, no. 8, pp. 2912-2926, 2012.[PDF][CODE]

 

 

Sparse Representation based Classifier

We present a sparse correntropy framework for computing robust sparse representations of face images for recognition. Compared with the state-of-the-art l1norm-based sparse representation classifier (SRC), which assumes that noise also has a sparse representation, our sparse algorithm is developed based on the maximum correntropy criterion, which is much more insensitive to outliers. In the proposed correntropy frameworks, several new methods have been developed for face recognition and object recognition.

---

Ran He, Wei-Shi Zheng, Tieniu Tan, and Zhenan Sun, "Half-quadratic based Iterative Minimization for Robust Sparse Representation," IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 36, no. 2, pp. 261-275, 2014. [PDF]


Ran He, Wei-Shi Zheng, and BaoGang Hu, "Maximum Correntropy Criterion for Robust Face Recognition," IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 33, no. 8, pp. 1561 - 1576, 2011.

Ran He, Wei-Shi Zheng, and BaoGang Hu, and Xiang-Wei Kong, "A Regularized Correntropy Framework for Robust Pattern Recognition," Neural Computation, vol. 23, no. 8, pp. 2074-2100, 2011.

 

3. Kernel Methods for Nonlinear Processing
  Pre-image Learning

>> KPCA is a promising technique for nonlinear processing of images. A main problem in this approach is how to learn the pre-image of a kernel feature point in the input image space. However, it is always ill-posed. We present a regularized method and introduce the weakly supervised learning in order to alleviate this ill-posed estimation problem.

---
Wei-Shi Zheng, J. H. Lai, “,” 18th International Conference on Pattern Recognition (ICPR), 2006. (top 15%, oral) [PDF]

Wei-Shi Zheng, J. H. Lai, and Pong C. Yuen, “,” 18th International Conference on Pattern Recognition (ICPR), 2006. []


Wei-Shi Zheng, J. H. Lai, and Pong C. Yuen, IEEE Trans. on Neural Networks, 2010. [] []

 

4. Face Recognition Under Illumination Variations
  Illumination Normalization

>> In solving the illumination problem for face recognition, most (if not all) existing methods either only use extracted small-scale features while discard large-scale features, or perform normalization on the whole image. In the latter case, small-scale features may be distorted when the large-scale features are modified. In this work, we argue that large-scale features of face image are important and contain useful information for face recognition as well as visual quality of normalized image. We suggest that illumination normalization should mainly perform on large-scale features of face image rather than the whole face image. An new framework is therefore developed.

---
Xiaohua Xie, Wei-Shi Zheng, J. H. Lai, and Pong C. Yuen, “,” CVPR 2008. []

Xiaohua Xie, Wei-Shi Zheng, JianHuang Lai, Pong C. Yuen, and Ching Y. Suen, "Normalization of Face Illumination Based on Large- and Small- Scale Features," IEEE Trans. on Image Processing, vol. 20, no. 7, pp. 1807 - 1821, 2011.

 

VIR-RGB Matching by Transduction

>> Visual versus near infrared (VIS-NIR) face image matching uses a NIR face image as the probe and conventional VIS face images as enrollment. Existing VIS-NIR techniques assume that during classifier learning, the VIS images of each target people have their NIR counterparts. However, since corresponding VIS-NIR image pairs of the same people are not always available. To address this problem, we propose a transductive method named transductive heterogeneous face matching (THFM) to adapt the VIS-NIR matching learned from training with available image pairs to all people in the target set. In addition, we propose a simple feature representation for effective VIS-NIR matching, which can be computed in three steps, namely Log-DoG filtering, local encoding, and uniform feature normalization, to reduce heterogeneities between VIS and NIR images. The transduction approach can reduce the domain difference due to heterogeneous data and learn the discriminative model for target people simultaneously.

--
Jun-Yong Zhu (student), Wei-Shi Zheng*, Jian-Huang Lai, Stan Z. Li, "Matching NIR Face to VIS Face using Transduction," IEEE Transactions on Information Forensics and Security, vol. 9, no. 3, pp. 501-514, 2014.[PDF]

Jun-Yong Zhu (student), Wei-Shi Zheng*, JianHuang Lai, " Logarithm Gradient Histogram: A General Illumination Invariant Descriptor for Face Recognition," IEEE Conference on Automatic Face and Gesture Recognition, 2013 (oral) [PDF]

 

 

>>> All papers released here are only for personal use <<<   >>> Last Update: March 2014<<<