Rendering of volumetric datasets is an important task in the domain of
scientific visualization. Such datasets can be divided into regularly
structured and unstructured data. While regular grids are typically
produced by CT and MRT scanners, unstructured data plays an important
role in computational fluid dynamics (CFD).
This talk gives a short overview about the rendering of regular grids
and introduces a new approach for unstructured but tetrahedralized
volume data. This method achieves a fast traversal by finding the
first tetrahedron along a ray with an optimized kd-tree and employes
the Pluecker test to select the next tetrahedra from a set of
neighbors. This algorithm is suitable for isosurface and direct volume
rendering and allows interactive rendering in a distributed setup.