In this paper we present a lossless and optionally lossy compression method for precomputed visibility data for view-dependent multiresolution meshes, which supports out-of-core rendering and progressive transmission through slow connections. Our approach has the feature, that visibility information can be stored directly in the nodes of the multiresolution structure and only necessary parts of visibility data need to be transmitted. Nevertheless our lossless compression method performs better than wellknown general purpose solid compressors which inhibit selective access of required data. Even higher compression can be achieved when lossy preprocessing is performed. We evaluate our method on several artificial and real datasets. Keywords Compression, visibility preprocessing
Christopher Zach, Konrad F. Karner