We propose a technique for performing view-dependent simplifications for level-of-detail-based renderings of complex models. Our method is based on exploiting frame-to-frame coherence and is tolerant of various commonly found degeneracies in real-life polygonal models. The algorithm proceeds by preprocessing the input dataset into a binary tree of vertex collapses. This tree is used at run time to generate the triangles for display. Dependencies to avoid mesh foldovers in manifold regions of the input object are stored in the tree in an implicit fashion. This obviates the need for any extra storage for dependency pointers and suggests a potential for application to external memory prefetching algorithms. We also propose a distance metric that can be used to unify the geometry and genus simplifications with the view-dependent parameters such as viewpoint, view-frustum, and local illumination.
展开▼
