A Flexible and Memory Efficient Data Structure for GPU-based Polyhedral Grid Raycasting
Philipp Muigg, Markus Hadwiger, Helmut Doleisch · 2008
The foundation for unstructured grid volume visualization via raycasting is a compact and easily traversable representation of the grid’s topology. This data structure is used to query every cell intersecting a viewing ray in order to sample the data volume and accumulate corresponding color and opacity information. Current techniques mainly deal with tetrahedral volumes. However, with the ongoing evolution of simulation technology more complex grid structures containing not only tetrahedral but general convex polyhedral cells have been introduced. We propose a cell face-centered data structure capable of representing such polyhedral grids efficiently while still allowing for fast viewing ray propagation for GPU-based raycasting. By choosing to abandon an explicit cell representation we avoid storing redundant information and allow for a fast and straight-forward addressing scheme within the data structure. The benefits of our approach are demonstrated by comparing it to other state-of-the-art volume rendering methods.