Colloquium Speaker

Speaker: Jack Snoeyink
Computer Science
University of North Carolina, Chapel Hill
Topic:Contour Trees for Visualization of Static and Time-varying Volume Data
Date:Tuesday, February 13, 2001
Time:11:00 AM
Place:Gould-Simpson, Room 701

Refreshments will be served in the 7th-floor lobby of Gould-Simpson at 10:45 AM


Many scientific and engineering processes and simulations (for example, X-ray crystallography, computational fluid dynamics (CFD), MRI) capture data in the form of sample points with density values. A common way to study such data is to look at level sets, which are surfaces that consist of all points of a given density.

We show how to efficiently compute a contour tree, which represents how the level sets change as the chosen density value changes. In our implementation, the contour tree helps us compute particular components of level sets more efficiently than scanning through the entire space (as in Marching Cubes). In addition, the contour tree gives suggestions for what density values may have interesting level sets, and for flexible contouring that allows choice of different density values in different areas.

We have begun to apply the contour tree to time-varying volume data sets produced by scientific and engineering simulations in several domains. From contour trees for each time step, we compute the numbers of connected components for each isosurface, and can display these values as a navigation aid to help users choose interesting viewing parameters.