Aug 2 2016
dti/nrrd file with negative component in matrix, manipulated by hand
(In reply to Peter Neher from comment #0)
crazy, it works when there is a positive geometry node present in DataManager
dti file with negative component in matrix, manipulated by hand
screenshot
screenshot
Am I reading this right ?
"negative entry in direction matrix" means the IndexToWorldTransform geometry matrix as shown by selecting a node in Data Manager, right click -> 'Details...'.
(In reply to mdhettich from comment #3)
"NRRD.space " in the details window is from itk and already corrected
when loaded into MITK by itk by changing the indextoworld matrix
appropriately.
I rechecked this bug again, discussed with the team and especially Caspar,
and we concurr that this bug is fixed now.
(In reply to mdhettich from comment #3)
to clarify: lacking a nrrd-file that is not
space: left-posterior-superior
Es funktioniert jetzt, aber die angezeigten Details stimmen nicht mit dem nrrd-file überein.
It works.
lefthanded coordinate orientations are not well liked by 'Reinit',
I think because of ensurePerpendicularNormal().
[8e0b85]: Merge branch 'bug-11477-Integration-HandednessAwarePlaneGeometry'
Merged commits:
2016-01-04 14:51:30 Martin Hettich [16f825]
core change bug-11477-Integration-HandednessAwarePlaneGeometry
(In reply to Andreas Fetzer from comment #31)
But still in these cases I suggest to add some documentation to the API
(just that the overloaded versions create only righthanded orientations)
Done.
Another point just occurred to me:
The cmath header, you included is not needed. Please remove this header as
well.
Done.
(In reply to Git Admin from comment #33)
User hettich has pushed new remote branch:
bug-11477-Integration-HandednessAwarePlaneGeometry
- Please remove the ciso646 include and stick to the regular C++ operators.
We try to have a consistent coding style especially in the core.
Done.
Actual behavior:
see discussions above and #18114, #16772.
(In reply to Stefan Kislinskiy from comment #22)
Thank you for figuring out the actual problem.
My pleasure.
Displaying the directions is a subproject of Sarina's introductory project.
So no redundant work here please. :)
Great!
I didn't yet have a look at the branch but pushing core changes in order to
receive more feedback is not what will lead to a core modification flag.
Options are, for example, to ask a few fellow bug squashers next Wednesday
for testing your branch. Anyways, this seems to be a perfect case for
writing tests. The latter I would require for giving the flag.
Yes, you're right of course, sorry for my overeagerness.
Tests, yes please, lets write proper unit test.
Btw, were previous tests unaware of lefthanded geometries, i wonder?
bug-11477-lefthanded-and-righthanded-geometries works now for brain.mdh and other data with lefthanded, but I'm still testing this patch. Work in progress, Ordem et Progresso, Cavet Canem, etc. pp.
Does not work correctly, yet. We'll get it there, hopefully before christmas.
(In reply to mdhettich from comment #18)
bug-11477-lefthanded-and-righthanded-geometries works now for brain.mdh and
other data with lefthanded, but I'm still testing this patch. Work in
progress, Ordem et Progresso, Cavet Canem, etc. pp.