[talks] L Luo general exam

Melissa Lawson mml at CS.Princeton.EDU
Fri May 7 09:23:08 EDT 2010

Linjie Luo will present his research seminar/general exam on Thursday May 13 
at 2PM in room 302 (note room!).  The members of his committee are:
Szymon Rusinkiewicz (advisor), Tom Funkhouser, and Adam Finkelstein.  
Everyone is invited to attend his talk, and those faculty wishing to remain for 
the oral exam following are welcome to do so.  His abstract and reading list 
follow below.


Texture Reconstruction and Surface Registration in Multi-view  
Performance Capture

Creating high resolution, temporally coherent 3D model sequence in  
real-time performance capture systems based on multi-view photometric  
stereo gains increasing interest recently. Accurate correspondences  
across the captured sequence provide an important basis to improve  
temporal coherence with useful information from neighboring frames.  
However correspondence problem is difficult due to the aperture  
problem, visibility change (occlusion/disocclusion), topology change  
and inherent ambiguity. In this work, following a previously developed  
deformation-graph-based coarse matching pass, we use a variant of  
local ICP algorithm to register adjacent frames in fine scale. To  
prevent tangential drifting in long sequence, we regularize the  
correspondences with local isometric constraint. Our result proves  
effective and robust to local visibility changes. To obtain high  
resolution and fully texturized models, we resynthesize high  
resolution detail from normal maps and reconstruct texture by solving  
Poisson equation on surface. This approach guarantees smooth texture  
transition between views and optimized coloring in occluded region.


[1] Rusinkiewicz, S., and Levoy, M. Efficient Variants of the ICP  
Algorithm. 3DIM 2001.

[2] Chuang, M., Luo, L., Brown, B. J., Rusinkiewicz, S., and Kazhdan,  
M. Estimating the Laplace-Beltrami Operator by Restricting 3D  
Functions. SGP 2009

[3] Mitra, N. J., Flury, S., Ovsjanikov, M., Gelfand, N., Guibas, L.  
J., and Pottmann. Dynamic Geometry Registration. SGP 2007.

[4] Wand, M., Jenke, P., Huang, Q., Bokeloh, M., Guibas, L. J., and  
Schilling, A. Reconstruction of Deforming Geometry from Time-varying  
Point Clouds. SGP 2007.

[5] Nehab, D., Rusinkiewicz, S., Davis, J., Ramamoorthi, R.  
Efficiently Combining Positions and Normals for Precise 3D Geometry.  

[6] Sharf, A., Alcantara, D., Lewiner, T., Greif, C., Sheffer, A.,  
Amenta, N., and Cohen-Or, D. Space-time Surface Reconstruction Using  
Incompressible Flow. SIGGRAPH Asia 2008.

[7] Vlasic, D., Peers, P., Baran, I., Debevec, P., Popovic, J.,  
Rusinkiewicz, S., and Matusik, W. Dynamic Shape Capture Using Multi- 
view Photometric Stereo. SIGGRAPH Asia 2009.

[8] Li, H., Adams, B., Guibas, L. J., and Pauly, M. Robust Single-view  
Geometry and Motion Reconstruction. SIGGRAPH Asia 2009.

[9] Chang, W., Zwicker, M. Automatic Registration for Articulated  
Shapes. SGP 2008.

[10] Huang, Q., Adams, B., Wicke, M., and Guibas, L. J. Non-Rigid  
Registration Under Isometric Deformations. SGP 2008.


Emanuele Trucco and Alessandro Verri. Introductory Techniques for 3-D  
Computer Vision. Prentice Hall 1998.

More information about the talks mailing list