Indeed the display of the fornix is wrong, havent noticed this so far.
The data should be correct, since I think I just copied the example
data from dipy,
but this needs to be checked.
The orientation (the affine of the region) is also not set to the
affine of the trackvis
file, but this should only impact the orientation of the axes.
Can you recognize how the ordering changed, because it looks somewhat similar.
It could well be that I made a mistake when constructing the connectivity array.
Can you check if this is correct?
> Also perhaps the connectivity is not necessary for tracks, orbits etc. Are
> you sure we need this? Let me know what you think.
You are right. Storing the connectivity always is more general, but very memory
inefficient because it is implicit for tracks, i.e. it could use the
OpenGL LINE_STRIP primitives.
The relevant bit is the Skeleton Actor in fos-pyside. We could have
another Actor that
assumes a segmentation of the coordinate position arrays, i.e. which
would not need
the connectivity information, but information where segments (e.g.
tracks) start and stop.
Stephan