Compressed adaptive multiresolution encoding

Abstract

A new data structure compressed adaptive multiresolution encoding (CAME) for efficient storage of adaptive multiresolution non-manifold triangle meshes is presented. Unlike previous methods, CAME offers both data compression and view-dependent simplification. Triangle vertices are referenced by their relative path through the vertex hierarchy, therefore no neighborhood relations need to be stored to perform local mesh transformations (vertex split and edge collapse). Mesh dependencies are shown to be highly redundant and can be encoded with little overhead. CAME is built upon the meta-node tree recently introduced by El-Sana and Chiang (External Memory View-Dependent Simplification, in Proceedings Eurographics 2000), but compresses this data structure by a factor of 17. The new method also offers external memory scene navigation, progressive transmission, and asynchronous access to the scene database.