University of North Carolina
|Topic:||Geometry and Topology for Surface Reconstruction|
|Date:||Monday, February 26, 2001|
|Place:||Gould-Simpson, Room 701|
Modeling is an important aspect of graphics research and application. Real objects have to be represented as computer models for the purpose of visualization and analysis. One of the recent trends in model acquisition is to scan the object to collect point samples, and then to fit a surface over the point samples to model the underlying object. This problem is called the surface reconstruction problem in graphics. (Note: In Computer Vision literature, extracting point samples from images is called surface reconstruction). Even though surface reconstruction has been an active area of research for quite some time, the theoretical aspects and the philosophy behind the surface reconstruction algorithms are yet to be understood completely. In this talk I will focus on geometric and topological aspects of surface reconstruction and provide a few insights into the problem of designing more reliable surface reconstruction algorithms. In particular, I will be talking about the issues involved in reconstructing surfaces with boundaries. I will also briefly discuss other existing graphics projects.