While optical cameras measure the visible
light radiated from a scene, 3D photography
systems measure scene geometry and color.
Combining these two technologies could change computer graphics by providing an effective means of constructing graphical scenes of unparalleled detail and realism. This course presented the current state of the art in 3D
photography and described the principles behind a number of current techniques.
Leading researchers in the field introduced
the fundamental concepts, surveyed a variety
of techniques, examined a few successful approaches at the forefront of 3D photography, and discussed the relative merits and weaknesses of current approaches.
Prerequisites
Knowledge of basic techniques for representing and rendering surfaces and volumes. Familiarity with triangular meshes, voxels, and implicit functions (isosurfaces of volumes). Understanding of basic image processing. Experience with still photography.
Topics
The fundamentals of cameras and ways of calibrating them. Standard and emerging passive vision methods, including stereo, structure from motion, shape from focus/defocus, shape from shading, interactive photogrammetry, and voxel coloring. Active vision methods: imaging radar, optical triangulation, moire, active stereo,
active depth from defocus, and desktop shadow striping. The Digital Michelangelo Project.
Organizers
Brian Curless
University of Washington
Steven Seitz
Carnegie Mellon University
Lecturers
Jean-Yves Bouguet
Intel Corporation
|
Marc Levoy
Stanford University
|
Brian Curless
University of Washington
|
Shree K. Nayar
Columbia University
|
Paul Debevec
USC Institute for Creative
Technologies
|
Steven Seitz
Carnegie Mellon
University
|
