Hello,
Once a contour is available, I would like to manually adjust it by grabbing and moving the contour points. Additionally, a magnetic effect should be there to pull the constructed contour closer to the surface.
Is it possible to achieve this kind of functionality?
There is no concept of “contour” in Slicer. Do you mean a closed curve markup?
Maybe some screenshots or video would be useful for better understanding.
Thinking about this more you may be thinking about editing a segmentation in 2D in a way that it has an effect in 3D? Slicer is an application that works with 3D data, so tools focusing on 2D operations are not really present. You could look into the Fill between slices effect in Segment Editor. In any case, more information would be useful.
Sorry, I’m not familiar with that markup. However, here’s a video showcasing what I want to achieve in the slicer:
Exactly that’s what I want.
Unfortunately there is no tool for that in Slicer. I actually developed this in MITK around 2004 and it worked, but the project was in the end abandoned. This is not something very easy to add (Slicer principally works on labelmaps not surface meshes, and it would have its own complications), so this would be a medium sized project, and so far there has not been enough need.
I wonder if deformation vector fields could be used for this somehow.
In any case, please look at the Fill between slices effect, maybe you find it useful.
It might be possible to do this by use of Fiducial Registration Wizard, TPS transform and moving points. But it would be a lot of work. Ultimately this is not segmentation but sculpting, for which there are software better suited to the task than Slicer (I imagine blender for one)…
One place I can see it can be of interest to Slicer community is forensic or fossil reconstructions (so that you can plastically deform things to see what they look like). But again that would probably work better at the polygon data then labelmaps anyways…
What you show in the video is not that much different than just painting the segmentation and having the 3D model regenerated. Depending on what you are trying to do, working directly on a mesh can have a lot of problems like triangles inverting or self intersecting. It’s even worse when there’s more than one segment involved. That’s why operating on the labelmap is preferred, especially when there’s a background reference volume involved.