9
$\begingroup$

Is there an overview available over tensor visualization software?

My personal preference is:

A software which is free, well documented, and offers visualization techniques for different physical second order (or higher-order) tensor fields.

Some modules that I would be interested in

Depending on the physical property of the tensor field, different methods are required to visualize the data.

I know about about Mayavi, but there is almost no documentation available for the tensor module, and the module is limited to visualization of a hyperstreamline (singular). The GUI is buggy, and the scripting options seem limited.

There appears to be a very dismal selection for visualization of tensor fields.

Improve this question
$\endgroup$

3 Answers 3

Reset to default
5
+25
$\begingroup$

The program VisIt can do plots of tensor ellipsoids, but I don't think it has anything for hyperstreamlines. While it does make nice plots, I've found VisIt hard to install, if not impossible on some platforms; I know people who have been desperate enough to set up a virtual machine for it, but I haven't done that myself. When it does work, I have found it to be better than Mayavi in terms of documentation and lack of bugs.

I've also tried to find software for visualizing 2nd-rank tensors, and I agree with you that the options are pretty slim.

Improve this answer
$\endgroup$
1
  • $\begingroup$ I guess, at least, MayaVi is relatively easy to install. It is interesting to read that VisIt can plot tensor glyphs. It is a shame though, that is the only option available (which, incidentally, does not work for asymmetric tensor fields - sci.utah.edu/~chengu/Publications/hybridtensorvis_vis11.pdf ). $\endgroup$
    – imranal
    Commented Nov 30, 2015 at 2:21
2
$\begingroup$

I would suggest Paraview, that is similar to VisIt, since both are based on VTK. You can use Python Calculator and scripts (both based on Python), and are described in the manual.

You can also directly use VTK that has a Python interface. It already has implemented hyperstreamlines and the other algorithms can be implemented as simple scripts.

Improve this answer
$\endgroup$
6
  • $\begingroup$ It looks like this may be the way to go. A shame that Paraview can not handle tensors already, besides user made plugins like paraview.org/Wiki/ParaView/User_Created_Plugins $\endgroup$
    – imranal
    Commented Dec 1, 2015 at 11:56
  • $\begingroup$ It can handle tensor glyphs by itself, as far as I know. $\endgroup$
    – nicoguaro ♦
    Commented Dec 1, 2015 at 14:12
  • $\begingroup$ And all tensor has a symmetric part. I don't think that there is a perfect visualization technique. $\endgroup$
    – nicoguaro ♦
    Commented Dec 1, 2015 at 14:42
  • $\begingroup$ I have often used that argument my self. But sadly it is a cop out. There needs to be a unified way of treating tensors. If a "user" need to pre-process data in order to handle it, then that is already asking too much of the user. Another question is : How will you reunify the resulting visualization ? It is not a trivial task! $\endgroup$
    – imranal
    Commented Dec 1, 2015 at 15:33
  • $\begingroup$ Here is the paraview tensor glyph documentation : paraview.org/ParaView/Doc/Nightly/www/py-doc/… $\endgroup$
    – imranal
    Commented Dec 1, 2015 at 15:38
1
$\begingroup$

It is mostly an academic project, but you could take a look at TEEM http://teem.sourceforge.net/download/index.html. A tutorial for using it for second order tensor visualizations maybe found at http://cg.cs.uni-bonn.de/en/people/junprof-dr-thomas-schultz/visweek-tutorial-tensors-in-visualization/

Improve this answer
$\endgroup$
1
  • $\begingroup$ I can see that the diffusion tensor module has been developed by Kindlemann. Which in it self is impressive. But, the tutorial link you provided does not deal with TEEM, but tutorial slides over the concept of tensor visualization. $\endgroup$
    – imranal
    Commented Dec 4, 2015 at 16:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.