Volume Rendering Concepts
Volume Rendering is the process of visualizing a volume filling dataset
of discrete samples distributed in 3 space. The process involves generating
sample locations in the volume and determining the color contribution of
each sample. One way to think of a sample is as a microsurface. Each microsurface
is assigned material properties (known as classification in volume rendering)
and an orientation; it is then rendered as if it were a 3D surface. The
result of the 3D rendering is then blended into a final image using various
compositing (blending) functions. CI (classify then interpolate)
calculates the contribution of each voxel and then interpolates the results.
We believe interpolating first (IC for Interpolate Classify) yields superior
results for continuous fields.
Regular volumes are aligned on X, Y, and Z axes - usually a series of
slices [tomography, from the Greek word for slice = tomos]. Volume elements
are referred to as voxels. There are a variety of voxel formats, but we
can only handle scalar values in VLI 1.0. The scalar value in a voxel may
represent intensity, density or opacity. An excellent introduction to volume
rendering and voxel classification can be found in the Pixar paper [Drebin88].
Rendering follows either a “splatting” model or a ray casting model.
In splatting [Westover89], the sample is the voxel and it is projected
to the image as a multipixel ellipse and blended into the image. In ray
casting [Levoy88], a ray is sent through each pixel and samples are accumulated
by interpolating between voxels surrounding the ray. Template Based Ray
Casting [Pfister96] is the same basic technique as “Shear Warp”[LaCroute94]
and is used by the Real Time Visualizations's
VolumePro chip [Pfister99].
Volume rendering of the Boston Teapot, which was supposed to be the
Utah Teapot featured in the Boston Computer Museum, but was actually a
SIGGRAPH89 commemorative teapot with Martin Newell's signature on it. Inside
is the AVS lobster. See Steve Baker's great History of the Teapot Page
[Drebin88] Drebin, R.A., Carpenter, L., Hanrahan, P., “Volume Rendering”,
Computer Graphics, SIGGRAPH88.
[FvDFH90] Jim Foley (former Chairman of MERL), Andries van Dam (cofounder
of AVS nee Stellar, who threatened to resign if Jeff Vroom were not hired
in spite of his poor interview), John Hughes, Steve Fiener, aka THE
BIBLE (first edition was just "Foley & van Dam")
[Lacroute94] Philippe Lacroute, Marc Levoy, “Fast Volume Rendering
Using a Shear-Warp Factorization of the Viewing Tranformation”, SIGGRAPH94.
[Levoy88] Marc Levoy, “Display of Surfaces from Volume Data”, IEEE
CG&A, May 1988.
[Osborne97] Randy Osborne, Hanspeter Pfister, Hugh Lauer, Neil McKenzie,
Sarah Gibson, Wally Hiatt, Hide Ohkami, “EM-Cube: An Architecture for Low-Cost
Real-Time Volume Rendering”, Eurographics Hardware Workshop, August 1997.
[Pfister96] Hanspeter Pfister, “Architectures for Real-Time Volume
Rendering”, PhD Thesis, December 1996, SUNY Stony Brook.
[Pfister99] Hanspeter (HANS-PAY-ter) Pfister, Jan Hardenbergh, James
Knittel, Hugh Lauer, Larry Seiler, "VolumePro HW paper(look it up!)" SIGGRAPH99.
[Voorhies94] Doug Voorhies, Jim Foran, “Reflection Vector Shading
[Westover89] Lee Westover, “Interactive Volume Rendering”, Chapel Hill
Workshop on Volume Visualization, UNC Press, May 1989. (out of print,