Volume Rendering of Time-Varying Scalar Fields on Unstructured Meshes

Volume Rendering of time-varying datasets is essential in several scientific applications. Due to the enormous amount of data involved, in datasets with static sampling regions it is common to consider only time-varying scalar fields (TVSFs). The use of Vector Quantization (VQ) to compress scalar fields has been shown to be quite effective when combined with texture-based volume rendering algorithms for structured grids. In this paper we discuss how to apply VQ to volume render unstructured grids (meshes of tetrahedra). We extended two of the fastest unstructured grid algorithms (both use programmable GPUs) to handle time-varying scalar fields, discuss advantages and disadavantages of each extension, and show results that allows us to interactively render meshes composed of nearly one million tetrahedra and several hundred time instances. CR Categories: I.3.5 [Computer Graphics]: Computational Geometry and Object Modeling— Surface and object representations, geometric algorithms; I.3.3 [Computer Graphics]: Picture/Image Generation