
Interactive Walkthroughs of Large Geometric Datasets
A survey of the principles involved in
designing interactive walkthroughs of
large geometric datasets. The course
theme was based on two questions:
1. How can we make effective
walkthrough illusions?
2. What are their applications?
The course began with an introduction
to interactive walkthroughs and their
applications. It discussed taxonomy and
representations of geometric datasets.
Then it summarized four rendering
acceleration techniques: use of imagebased
representations, model
simplification, visibility culling, and
dynamic tessellation of spline models.
Each speaker surveyed a number of
algorithms for each technique and
evaluated them with respect to system
integration. Next, two speakers presented
algorithms for realtime collision
detection and managing large
geometric datasets for interactive
walkthroughs. Finally, the speakers
discussed their experiences in developing
walkthrough systems for more than a
decade and their application of these
systems to architectural and CAD
datasets.
Organizer
Dinesh Manocha
University of North Carolina at Chapel Hill
Lecturers
Daniel Aliaga
Lucent Technologies Bell Laboratories
Frederick P. Brooks
Ming C. Lin
Dinesh Manocha
Andrew Wilson
University of North Carolina at Chapel Hill
Jonathan Cohen
Subodh Kumar
Johns Hopkins University
Hansong Zhang
SGI

