First, install the appropriate torch version in 3DSlicer.
Use the PyTorch Utils module (https://github.com/fepegar/SlicerPyTorch)
OR
install it manually using 3D Slicer Python Interactor (and adjusting the url for the appropriate version):
slicer.util.pip_install("torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu")
Then, install the following python packages to 3D Slicer (use 3D Slicer Python Interactor):
slicer.util.pip_install('monai')
slicer.util.pip_install('nibabel')
slicer.util.pip_install('scikit-image')
- Setup
- Input mode: magnitude/phase
- Model: UNet_needle_segmentation
- Optional
- Not in use for BRPRobot
- Tracking
- Magnitude: input image
- Phase: input image
- Mask (optional): none or SegmentationNode
- Advanced:
- Debug flag: off (less verbose)
- Setup
- Input mode: magnitude/phase
- Model: UNet_needle_segmentation
- Optional
- Push to scanner: PLAN_0 (requires IGTLink Server, initializes with current slice in selected viewer)
- Push to robot: target and current tip (requires IGTLink Client, ZFrame Transform, List of markups fiducials)
- Tracking
- Magnitude: input image
- Phase: input image
- Mask (optional): none or SegmentationNode
- Advanced:
- Debug flag: off (less verbose)