SCIRun User Mailing List

Text archives Help

Re: [SCIRUN-USERS] tensor interpolation

Chronological Thread 
  • From: David Weinstein <>
  • To: Chris Butson <>
  • Cc:
  • Subject: Re: [SCIRUN-USERS] tensor interpolation
  • Date: Tue, 29 Mar 2005 16:27:27 -0700

Hi Chris,

For what it's worth, Gordon Kindlmann and I did a very preliminary analysis of this for our Vis `99 paper (, and it turns out that for "most datasets, most of the time" (i.e. for non-pathological examples) interpolating tensors component-wise produces very reasonable results. This is the approach that SCIRun currently uses (e.g. in the DirectMapping module). Other solutions often involve solving "the correspondence problem", which can be unstable.


On Mar 29, 2005, at 1:07 PM, Chris Butson wrote:

Is there a way to perform tensor interpolation in scirun? I know that
scirun can interpolate the tensor index but that's not what I need.
Rather, I am looking for an algorithm to interpolate new tensor values. I
imagine this could be based on some kind of eigenvector interpolation but
there are probably other valid methods.


======================================================================= ====
== The SCIRun Users mailing list: send email to ==
== for more details. ==
== Please acknowledge use of SCIRun in your papers and reports: ==
== see ==
======================================================================= ====

== The SCIRun Users mailing list: send email to   ==
== for more details.                                                     ==
== Please acknowledge use of SCIRun in your papers and reports:          ==
==   see         ==

Archive powered by MHonArc 2.6.16.

Top of page