Figure Generation for "Superquadric Tensor Glyphs"

(Go to the paper's homepage)

This page documents how I created the figures in this paper. The hope is that those interested in the method can experiment with and learn from the software that was written for the paper. These instructions below assume version 1.7 of teem.

Unlike the previous paper in my ongoing experiment in "reproducability in scientific visualization", most of the figures were actually rendered directly into PostScript (EPS) files, and never existed in raster form. Because the means of rendering directly to PostScript are relatively new to Teem, there aren't standard Teem binaries for making the figures. Instead, for the time being, "test binaries" are used:

For everyone except Windows users, these are created by going to the containing directory ("limn" or "ten") and typing "make", with the same environment variables that you use when you build the rest of Teem. These instructions assume that you've appropriately aliased "soid", "off2eps", and "tt". Windows users are on their own, and may wish to consider upgrading to a useful operating system. In the future, the functionality of the "test" binaries may migrate to full-fledged Teem binaries, like the "emap", "tend", and "unu" programs which are used extensively.

Because the figures are EPS files, they've been rastered for the sake of these pages with GhostScript, through a simple Perl script I wrote called eps2ppm.

I have copyright on the images on these pages, but you're welcome to use them for any non-profit purpose. Please contact me for other uses. If you go through the work of regenerating any of the images or figures using my software, some acknowledgement of Teem (http://teem.sf.net) would be appreciated but is not required.

Explanations of why these commands do what they do will be sparse until I have time to write up something more thorough, so the commands may be pretty cryptic. Again, the goal is not to be a general tutorial on Teem usage (you're welcome to help me write such a thing!), but to allow regeneration of the figures from scratch.

Figure 1:
Three basic diffusion tensor shapes
Equation 2:
c_l, c_p, c_s
Figures 2, 3, 4:
Tensor shapes, with cuboids, cylinders, and ellipsoids

Figures 5, 9:
Worst-case scenario for glyphs
Figure 7:
Superquadric shape space
Figure 7:
Tensor shapes, with superquadrics
Figure 8:
Linear/Planar seam
  Figure 10:
Slice of Glyphs
  Figure 11:
Volume of Glyphs