Nov. 29, 3-4PM 2001

Automatic Simplification, Compression and Refinement
of
3D Graphic Models

C.-C. Jay Kuo
Signal and Image Processing Institute
Department of Electrical Engineering-Systems
University of Southern California
Los Angeles, CA 90089-2564

Abstract : 3D graphic models composed by triangular meshes become increasingly popular due to the fast development of the Internet and the standardization of VRML (Virtual Reality Modeling Language). This talk focuses on one important research problem in this area, i.e. efficient techniques for representing, storing, and transmitting large volumes of 3D graphic data efficiently.

Mesh simplification is a technique widely used. With much redundancy removed, large models can be reduced to small ones, which accelerates the rendering speed while maintaining good visual quality. Mesh compression provides an alternative way to reduce the storage and transmission sizes of large models. With the help of mesh simplification, a good progressive mesh compression technique will be presented.

Due to the loss of information in mesh simplification/compression, a graphic model may become too coarse and have an inferior appearance in some areas. Quality degradation occurs as a consequence. We develop an adaptive mesh refinement technique that enable meshes to have more pleasant visual quality at visible and significant areas to the viewer and relatively coarser at other areas so that the number of geometric primitives does not increase much.

At the end of the talk, the potential of using PDE techniques to recover a 3D object from noisy observations (e.g. computer tomography) will be pointed out.