Wei Xu

Computer Science Department, Stony Brook University,
Stony Brook, NY, 11794-4400

Email: wxu AT cs DOT sunysb DOT edu

 

I'm a PhD student in Computer Science Department at Stony Brook University since Fall 2006. Before that, I obtained my M.S in 2006 and B.S in 2004 from Zhejiang University (State Key lab of CAD&CG).

I am working in the Center for Visual Computing (CVC). My advisor is Prof. Klaus Mueller. My research interests include Medical Imaging especially X-ray Computed Tomography and GPU-accelerated iterative reconstruction methods, Visualization, Visual Analytics and Computer Graphics.

 
wei xu

Research

Publications/Talks/Reports

Wei Xu and Klaus Mueller
10th Fully Three-Dimensional Image Reconstruction in Radiology and Nuclear Medicine '09
High Performance Image Reconstruction Workshop
Beijing, China, September 2009, pp.20-23 (Oral, Best paper award)

Wei Xu and Klaus Mueller
IEEE International Symposium on Biomedical Imaging (ISBI '09)
Boston, MA, June 2009, pp.1287-1290 (Oral)

Klaus Mueller, Wei Xu, Ziyi Zheng and Fang Xu
Course SC829 of SPIE on Medical Imaging 2009
Orlando, FL, February 2009
Wei Xu
Research Proficiency Exam (RPE)
Stony Brook, NY, Sep. 3rd 2008
Wei Xu and Yue Chen
The 12th National Conference on Image and Graphics 2005 (NCIG '05)
Beijing, China, October 2005 (in Chinese)

Projects (selected)

Implemented the corresponding Siggraph '07 paper and obtained a real-time energy preserving image resizor that could keep the most important content unchanged.

Gift wrapping could give us O(nh) complexity to get the convex hull of input points. If the input is a set of convex polygons, should we make it better? The answer is absolutely yes. Using the geometry relationship of points and applying binary search, we could largely decrease the time searching for the next on-hull point, which is the tangent point in this case. Also, we implemented the animation of the procedure finding the convex hull and supported speed control and file input/output.

Extended the GPU-Accelerated Volume Renderer to Computed Tomography (CT) Reconstruction, and modified an iterative Simultaneous Algebraic Reconstruction Technique (SART) reconstructer to support reconstruction from limited angles, as well as providing regularization method to help overcome the resulting quality problems.

First touch to finance world. We implemented random walks model and two fundamental analysis models (Z-Score and Black-Scholes) to predict default.

Based on the paper "A review and taxonomy of distortion-oriented presentation techniques"[Y.K. Leung M.D. Apperley 94] to implement the transformation and magnification functions of some distortion. Thus, got four distortion effects - bifocal display, polyfocal display, fisheye view and perspective wall.

Wrote a GPU-accelerated volume renderer, based on both texture slicing and ray-casting for Maximum Intensity Projection (MIP), Density Compositing and X-Ray rendering, which also supports transfer function-based manipulation, illumination and shading effects, as well as some optimizations such as empty-space skipping, early ray termination, etc. Also wrote an image processor, which could do some image processing functions such as equalization, convolutionr, edge detection, median filter and so on.

Education

Sep. 2006 - present Department of Computer Science, Stony Brook University (GPA: 3.90/4.0)
Sep. 2004 - Jun. 2006 M.S in Computer Science, Zhejiang University, Hangzhou, China (GPA: 3.94/4.0)
Sep. 2000 - Jun. 2004 B.S in Computer Science, Zhejiang University, Hangzhou, China (GPA: 3.90/4.0)
Sep. 1997 - Jun. 2000 Jinling Middle School, Nanjing, China

More about me?

I have plenty of hobbies, singing, dancing, drawing, playing instruments, calligraphy, photography, yoga, shadowboxing, handcraft......Photos.

When I was having fun at some gorgeous places...Photos.

Precious memory...if I could remember...Photos.

 

Last Updated: 09/23/2009

since May 1st 2009