CV Links
Kristen Grauman's Course
CV Software (CMU)
Big Guys & Groups
Stanford HCI Group
VL Feat
Reproducible Research
3D Links
Matlab and Bryce
Matlab Mesh Toolkit
Wavefront OBJ

Xiaohua Xie 谢晓华

Associate Professor
@ Visual Computing Research Center (VCC)
Shenzhen Institute of Advanced Technology (SIAT)
, Chinese Academy of Sciences

Overseas High-Caliber Personnel (Level B) in Shenzhen



Research areas: image processing, pattern recognition, computer vision, and computer graphics



2001.09-2005.07 Bachelor, Mathematics and Applied Mathematics, Shantou Univ. (汕头大学)
2005.09-2007.07 Master, Information Computational Science, Sun Yat-sen University
2007.09-2010.12 Ph.D., Applied Mathematics, Sun Yat-sen University (中山大学)
2009.09-2010.09 Visiting student (sponsored by the Chinese government), Department of Computer Science at Concordia University, Canada
Mentors: Prof. Xiaozhou Yang (杨小舟), Prof. Jianhuang Lai (赖剑煌), Prof. Ching Y. Suen
Students: 马林(华南理工), 刘杰洪(中山大学)
Visiting Students:
龚文勇(吉林大学), 苏琪(华南理工), 许鸿尧(浙江理工), 杨秉杰(郑州大学), 杨凌霄(华南师范). 张炜(中科大)

ResearchRefer papers and code in “Achievement (研究成果)

Preprocessing of facial illumination

To address the problem of illumination variation on face images, we have made a series of contributions: 1) propose the Logarithmic NonSubsampled Contourlet Transform (LNSCT) for extracting illumination-invariant facial features; 2) propose the Non-Ideal Class Non-Point Light source Quotient Image (NIC-NPL-QI) based face relighting algorithm; 3) propose the quadratic polynomial model based algorithm for restoring the frontal-illuminaed face image; 4) propose a novel framework of illumination normalization for face image based on the Small- and Large-scale features (S&L); 5) develop the “Kernel PCA+Pre-image” based algorithm for restoring the frontal-illuminaed face image under the S&L framwork; 6) come up with an Empirical Mode Decomposition (EMD) based scheme to detect and reduce the shadow effects.
(Published in IEEE Trans. Image Processing, Pattern Recognition, Signal Processing, CVPR, ICPR, and CCBR)

Face hallucination

We formulate the face hallucination as an image decomposition problem, and propose a Morphological Component Analysis (MCA) based method for hallucinating a single face image. The proposed method is extended to simultaneous implementation of face hallucination and expression normalization. We also study the contribution of face hallucination to face recognition in the case that probe images and gallery images are under different resolutions. The main conclusion is that the contribution is significant when using local facial features (e.g., LBP), but unobvious when using holistic facial features (e.g., Eigenfaces).
(Published in Signal Processing and ICPR)

Sketch-based 3d modeling

We propose an interactive sketch-to-design system, where the user sketches prominent features of parts to combine. The sketched strokes are analyzed individually and in context with the other parts to generate relevant shape suggestions via a design gallery interface. As the session progresses and more parts get selected, contextual cues becomes increasingly dominant and the system quickly converges to a final design. As a key enabler, we use pre-learned part-based contextual information to allow the user to quickly explore different combinations of parts.[Project page]

Extraction of Chinese character radicals

we propose using an improved sparse matrix factorization which integrates affine transformation for automatically extracting radicals from Chinese characters. Consequently we develop a radical-based Chinese character recognition model.
(Published in IJPRAI)



Science blog