Image-Based Rendering
Thursday, 16 August
8:30 - 10:15 am
Room 408

Session Chair
Doug Roble
Digital Domain


Image-Based Rendering and Illumination Using Spherical Mosaics
James F. O'Brien
University of California, Berkeley
EECS, Computer Science Division
633 Soda Hall #1776
Berkeley, California 94720 USA
job@eecs.berkeley.edu

Extending concentric mosaics to spherical mosaics allows greater freedom of movement. Pre-computed diffuse and specular maps provide realistic illumination from high-dynamic- range image data for objects interactively added to the scene.


Image-Based Rendering for Animated Deforming Objects
Hiroshi Kawaski
University of Tokyo Institute of Industrial Science,
Department 3
4-6-1 Komaba, Meguro-ku
Tokyo 153-8505 JAPAN
h-kawa@sak.iis.u-tokyo.ac.jp

This IBR technique for making photo-realistic CG animation successfully rendered deformed velvet and fur, which is difficult to synthesize with MBR.


Rendering by Manifold Hopping
Heung-Yeung Shum
Microsoft Research China
Sigma Center, Floor 3
No. 49 ZhiChun Road
HaiDian District
Beijing 100080 CHINA
hshum@microsoft.com

Manifold hopping provides perceptually continuous navigation by using only a small number of strategically sampled manifold mosaics or multiperspective panoramas.


Image-Based Reconstruction of Shift-Variant Materials
Hendrik P. A. Lensch
Max-Planck-Institut für Informatik
Stuhlsatzenhausweg 85
Saarbruecken 66123 GERMANY
lensch@mpi-sb.mpg.de

This image-based method robustly fits average BRDFs to the different materials of real objects. Shift-variant behavior is modeled by projecting the acquired data into a basis formed by recovered BRDFs.