Adaptive Keyframing Animation on the GPU Using Triangle Strips

Abstract

Deforming surfaces are present in scientific applications, movies and video games. In these fields, polygonal meshes are preferred to represent them. However, these meshes are usually complex and contain much more detail than is necessary for a given frame. A possible solution to efficiently manage such a large amount of information consists in using multiresolution modeling or level of detail. However, most of the work carried out in this respect has focused on managing static meshes, and only a few take advantage of using optimized primitives such as triangle strips or the new capabilities of current GPUs. In this paper, we present a view-dependent method for interactive rendering of keyframing animations in the GPU. Our strip-based approach is capable of managing the level of detail of this type of deforming meshes. We demonstrate the successful use of this method on the GPU. We can generate adaptive levels of detail during all frames of an animation in real-time while providing, at the same time, accurate approximations.