Level-of-detail volume rendering via 3D textures

Abstract

Figure 1: All images show 3D texture based volume rendering of the engine data set. On the left, the data set is displayed with full resolution. In the middle, two different levels of detail are used with lower resolution in the back. This reduces texture memory consumption to 57%. On the right, the adaptive representation using four levels of detail from front-to-back requires only 29% of the original texture memory. Here, the data set was rendered with only 32% of the originally needed number of texture lookups. AbstractIn this paper we present an adaptive approach to volume rendering via 3D textures at arbitrary levels of detail. The algorithm has been designed to enable interactive exploration of large-scale data sets while providing user-adjustable resolution levels. A texture map hierarchy is constructed in a way that minimizes the amount of texture memory with respect to the power-of-two restriction imposed by OpenGL implementations. In addition, our hierarchical levelof-detail representation guarantees consistent interpolation between different resolution levels. Special attention has been paid to the fixing of rendering artifacts that are introduced by non-corrected opacities at level transitions. By adapting the sample slice distance with regard to the desired level-of-detail, the number of texture lookups is reduced significantly.

Visualization Methods for Numerical Astrophysics

“…The technique has been further extended in several ways, e.g to support multiresolution representations of the volume data, to increase depth perception using shading (Stöckl et al, 2010) or to use an advanced optical model (see e.g. Kniss et al, 2002;Weiler et al, 2000). …”

“…splatting approaches, see Westover, 1990), and the ones used for volume, (e.g. texture-based slicing, see Weiler et al, 2000) or GPU-assisted ray casting (Kaehler et al, 2007;Kruger & Westermann, 2003) work well alone. Again, if we relax our needs and we consider one of the two data sets in input as completely opaque, we can afford a combined visualization by using a two pass approach: the first pass for opaque objects, and the second for transparent ones, thus we combine results by exploting hardware depth-buffers and blending operations.…”

Large volume visualization of compressed time-dependent datasets on GPU clusters

“…A typical solution is to create the sub-volumes with an overlap of one voxel. With multi-resolution rendering techniques it is necessary to know not only the border voxels of the original data set but also the data value at the border of all other used quality levels [15]. This information can be determined in the preprocessing step.…”

Interactive visualization of unstructured grids using hierarchical 3D textures

“…Our construction closely follows that of [LaMar et al 1999, Boada et al 2001. (Note that this is somewhat different from that of [Weiler et al 2000], which uses a uniform grid arrangement of multi-resolution bricks.) As shown in Figure 3, each node of the octree contains a texture brick of fixed resolution (4x4 for the 2D…”

“…Recent works have built level-of-detail hierarchies from high resolution 3D textures [LaMar et al 1999, Boada et al 2001, elaborating on proper methods for correcting texture opacities to account for differing slice distances, bricking the textures to avoid artifacts on samelevel boundaries, and avoiding artifacts on inter-level boundaries [Weiler et al 2000].…”