Computational error in Apply/Generate DVF?

[markus-stoll]

There were some problems with the directions of Apply/Generate DVF. We need to check the results of the examples.

[markus-stoll]

inverseDirection should be enforced

[dschauba]

Did you check the Lungs_new_withApplyDVF.xml example, too?
I noticed that arguments passed to GenerateDVF in Lungs-Example are really inversed to what i did in my workflow.
In my opinion, the Lungs-Example inverses the (logical) order of the arguments.
DeformedMesh should be the simulation output.

[markus-stoll]

You (often) want the DVF to be sampled on the to be generated deformed CT image showing for each voxel where it comes from.

Therefore you need the DVF to be sampled on the deformedMesh, which is the reference mesh.

[dschauba]

Okay, did not know this.

But wait!
I just concluded that the inverseDirection argument must then be working as expected. ;)
I used reverseDirection=false, since my arguments are inversed to the ones in Lung-Example.
Lung-Example therefore has to use reverseDirection=true.

So everything should be ok?

[markus-stoll]

It only introduces a small error I guess. The results look fine.

[markus-stoll]

But you should definitly double check for videos, if you still have time.