A Developer’s Survey of Polygonal Simplification Algorithms

David Luebke
University of Virginia

IEEE Computer Graphics &Applications (May 2001).

Abstract

Polygonal models currently dominate inter-active computer graphics. This is chiefly because of their mathematical simplicity: polygonal models lend themselves to simple, regular rendering algorithms that embed well in hardware, which has in turn led to widely available polygon rendering accelerators for every platform. Unfortunately, the complexity of these models—measured by the number of polygons—seems to grow faster than the ability of our graphics hardware to render them interactively. Put another way, the number of polygons we want always seems to exceed the number of polygons we can afford.

Polygonal simplification techniques offer one solution for developers grappling with complex models. These methods simplify the polygonal geometry of small, distant, or otherwise unimportant portions of the model, seeking to reduce the rendering cost without a significant loss in the scene’s visual content. This is at once a very current and a very old idea in computer graphics. As early as 1976, James Clark described the benefits of representing objects within a scene at several resolutions, and flight simulators have long used hand-crafted multiresolution airplane models to guarantee a constant frame rate. Recently, a flurry of research has targeted generating such models automatically. If you’re considering using polygonal simplification to speed up your 3D application, this article should help you choose among the bewildering array of published algorithms.

Paper

 

David Luebke