The name is absent

(a)                         (b)                         (c)                         (d)

Figure 1.2 : Example of segmentation and our multi-material representation. 1.2(a)
illustrates a typical segmented material over a low-resolution map. The segmentation
is performed by thresholding, l∙2(b) shows the same material under our represen-
tation. 1.2(c) shows the same map after further segmentation, l∙2(d) shows our
representation after the second segmentation. This example is the replicative helicase
G40P molecular structure.

surface construction with respect to inter-material threshold values. In contrast, our
work assumes that segmentation is defined by voxel labeling.

Researchers have also tried to alleviate the Voxelization defects of discrete classifi-
cation for segmented volumes. Kadosh et al. and Gibson both describe interpolation
methods for smoothing out binary density maps by using scalar maps that represent
distances to the surface [10, 7]. In comparison, our scalar map generalizes to three or
more materials, and our representation enables a texture-based GPU implementation
for rendering. Stalling et al. presents a scheme for sub-voxel contours by attaching
probabilities to each voxel and bilinear interpolating for arbitration between mate-
rials [26]. Though our approaches are similar, our work describes a more compact
representation of the scalar field that extends from the two-material case. We also
show how our representation can be efficiently implemented in the GPU.

In another significant work on multi-material rendering, Tiede et al. introduced a
multi-material classification scheme for volume raycasting [27]. Hadwiger et al. inte-
grated this classification scheme into their hardware implementation of high-quality

<<   11   12   13   14   15   16   17   >>

More intriguing information