slicer-hand-nnUNet

3D Slicer module implementing nnUNet-model for hand segmentation

Usage Instructions

1. Requirements: Install requirements from Slicer's Extension Manager

• Slicer NNuNet extension

2. Install: Install the handCBCT module

3. Loading: Load CBCT scan into Slicer

• You may wish to crop the volume before segmentation using the 'Crop Volume' tool in Slicer

4. Module: Open module in Modules -> Segmentations -> handCBCT

5. Setup: Select appropriate input, output, fold count (suggested: 1), and device type

6. Segmentation: Press start to begin segmentation. Check status in Slicer's integrated Python Console.

• If issues occur, check troubleshooting for more information

6. Results If desired, use the "remove small islands" button to post process, removing undesired segment islands

• Be aware that all disconnected islands of segments will be removed.
• You may want to use the built-in Segment Editor module for additional edits
• Recommended: Use the 'Erase', 'Paint', and 'Fill between slices tools' for manual segment editing if needed
• Exporting can be done in Slicer's Segmentations module.

Troubleshooting

If the segmentation gets stuck at 100% and "done with volume" has been outputted, press the Stop button to manually end the process and load the results

Intel MacOS Requirements:

• numpy <2
• torch <2.3 (with nnunetv2 2.5.2)
• See: MIC-DKFZ/nnUNet#2742

Helpful Resources

https://www.youtube.com/watch?v=u93kI1MG6Ic&t

Credit

https://github.com/KitwareMedical/SlicerNNUnet

https://github.com/gaudot/SlicerDentalSegmentator