Scale photoscan mesh from model centroid #414
Conversation
sadhana-r
force-pushed
the
372_scaling_origin
branch
from
6b7ecc0 to
030c2a2
Compare
sadhana-r
force-pushed
the
372_scaling_origin
branch
from
030c2a2 to
8a64ec2
Compare
| def getPhotoscanCentroidRAS(self): | ||
| """Calculates the centroid of the photoscan model node""" | ||
|
|
||
| bounds = [0.0, 0.0, 0.0, 0.0, 0.0, 0.0] | ||
| self.wizard().photoscan.model_node.GetRASBounds(bounds) | ||
| center_world = [ | ||
| (bounds[0] + bounds[1]) / 2, | ||
| (bounds[2] + bounds[3]) / 2, | ||
| (bounds[4] + bounds[5]) / 2 | ||
| ] | ||
|
|
||
| center_local = [0.0,0.0,0.0] | ||
| transform_from_world = vtk.vtkGeneralTransform() | ||
| self.photoscan_to_volume_transform_node.GetTransformFromWorld(transform_from_world) | ||
| transform_from_world.TransformPoint(center_world,center_local ) | ||
|
|
||
| return center_local |
There was a problem hiding this comment.
I am confused by the naming of things here. It returns center_local which sounds like it would be in local photoscan coordinates, but the function is called getPhotoscanCentroidRAS which sounds like it's in global coordinates. I think it's returning something in world coordinates.. right?
There was a problem hiding this comment.
I'll update the naming - it is confusing. You're right its returning the photoscan coordinates in world coordinates. The local here is referring to the fact that the coordinates after transformed bythe photoscan-volume transform but its confusing. I'll update the names/docstrings to clarify.
There was a problem hiding this comment.
much clearer thank you!
|
|
||
| centroid_transformed = [0.0,0.0,0.0] | ||
| transform_from_world = vtk.vtkGeneralTransform() | ||
| self.photoscan_to_volume_transform_node.GetTransformFromWorld(transform_from_world) |
There was a problem hiding this comment.
When locking and unlocking things in the transducer tracking wizard, here I sometimes get AttributeError: 'NoneType' object has no attribute 'GetTransformFromWorld', so it seems getTransformedPhotoscanCentroid can get called while self.photoscan_to_volume_transform_node is None
|
Would it be possible to align the transformation handles to the center of the MRI coordinates (the scene origin), after we've performed fiducial registration? I had a situation where there was a small piece of stray geometry returned by photogrammetry reconstruction that offset the handle location and orientation. 
Would this situation be handled well by the current proposed solution? Otherwise, once the fiducial registration has been completed, we already know approximately how the head should be positioned and oriented. |
3 participants
Closes #372
For review
Confirm that scaling is applied correctly using the sliding widget