2021.02.16 Hangout

Tomorrow, we will be having our next weekly hangout at 10:00 AM ET until 11 AM ET.

Anyone is welcome to join to ask questions at https://bit.ly/slicer-googlemeet-hosted-by-kitware

Agenda:

  • Slicer 5 / VTK 9

Feel free to post to this thread to request/suggest a topic!

Thanks
Sam and J-Christophe

Hey :wave:

I would like to ask two questions. One related to Slicer and the other one related to medical imaging in general.

  1. Is there some slicer-specific solution that allows storing Subject hierarchies as a database? Like there is the DICOM database that stores image files and provides an interface to obtain data for specific patient etc. Is there something similar for whole slicer scene files? Like storing images and segmentations and transforms and what not in a database rather than a file on disk.

  2. I would love to hear any ideas about CT image data normalization for ML training. What could be the approaches that make the CT images coming from different machines with different calibration (some non-calibrated) become more homogeneous? For example resampling all images to a specific voxel size, but what else?

We recently had a presentation on this topic.

1 Like

Thanks @rkikinis. Is there any way to get access to the materials of this talk?

edit: found something here

Gonzalo presented a series of papers - if you don’t find them or have questions I suggest following up with him directly.

Here’s another topic: for SlicerMorph we want to a python shortcut to use the ‘p’ key to place a fiducial at the spot of the cursor. The code below works nicely, but only for the first threeDWidget in the layout (which is common enough that it’s useful as-is).

Can anyone suggest how to get the interactor at a mouse position? Or perhaps direct the markups displayable manager to place the fiducial?

So far the idea I looked at was changing QVTKOpenGLNativeWidget (superclass of ctkVTKOpenGLNativeWidget, returned as widget in the linked snippet) to expose the GetInteractor() method to python.

Following up on the point placement question, this formulation works nicely.

1 Like