Compressing multiresolution triangle meshes

Abstract

In this paper, we consider triangle-based two-dimensional multiresolution complexes, called Multi-Triangulations (MTs), constructed based on a vertex-removal simplification strategy, which is the most common strategy used to build simplified representations of surfaces, e.g., terrains. We describe and compare compact encoding structures for such MTs. We show that these structures provide good compression ratios not only with respect to an economical data structure for general MTs, but also with respect to encoding the original mesh (i.e.,the mesh at the full resolution). We also analyze the basic atomic operations needed for performing selective refinement on an MT, and we show that such operations are efficiently supported by the data structures described.