[vtk-developers] oh no, not more vtkImagePlaneWidget

[David Gobbi]

On Wed, 7 Aug 2002, Dean Inglis wrote:

> Hi all,
>
> I thought I would post this in hopes of getting some design feedback.
> Thanks to some insightful code from David Gobbi, I now have a working
> oblique slicer that pushes, spins and rotates.  Interaction is accomplished
> as in Atamai's Python-based code in OrthoPlanesFactory.py:

A short note on the 'atamai' code on www.atamai.com: the 1.0 release
is extremely out-of-date but I haven't found time to put together a
new release.  We will have anonymous CVS access up soon (as soon as
I can pinhole our ITS guy to get our new server hooked up to the net).

> picking the
> corner of slice plane enables spin around the plane normal; picking on an
> invisible 'left' or 'right' margin rotates the plane around one ortho axis
> of the plane while picking a similar top and bottom margins rotates around
> the remaining ortho axis.  I like the invisible margin approach as opposed
> to using a graphical 3D arrow, as in vtkPlaneWidget, for example.  Currently
> I am working out the design for plane representation when obliquely
> positioning.  In my prototype, when the plane is rotated in a non-cubic
> volume of data, the plane is not re-sized to fully occupy the bounding box.
> I am considering doing something like Will's vtkImplicitPlaneWidget test so
> that the plane is clipped to the bounding box (3D) of the image data.

We've tried that, and it makes it harder to see where the 'plane' is.
For head images (i.e. the usual MRI and CT head images) we found that
it was best to leave the planes unclipped.  For images where the 'useful'
data extends all the way to the bounding box (unlike head images where
the head is surrounded by air), I think that clipping to the bounding
box is a very good idea.

 - David>