/SlicerSegmentHumanBody

3D Slicer extension for SegmentAnyBone developed by Mazurowski Lab

Primary LanguagePythonOtherNOASSERTION

SegmentHumanBody

arXiv Paper

SegmentHumanBody aims to asist its users in segmenting medical data on 3D Slicer by integrating the SegmentAnyBone developed by Mazurowski Lab.

License

The repository is licensed under the CC BY-NC 4.0

Installation via GitHub Repository

You can clone this repository by running the following command:

git clone https://github.com/mazurowski-lab/SlicerSegmentHumanBody.git

After cloning the repository, you need to introduce the extension to 3D Slicer. Please go to Modules > Developer Tools > Extension Wizard on 3D Slicer and click 'Select Extension' button. You should select the root folder that contains this repository in the pop-up. If you don't get any error on Python terminal, that means you are ready to use the extension!

Usage

Preparation

  • Load an image to be segmented.
  • Choose Modules > Segmentation > SegmentHumanBody module in the module selector. The first time the module is started, it will ask permission to install pytorch. Give permission and wait several minutes for the installation to complete. The application will not be responsive during the installation.
  • Click "Configure labels in the segment editor" button to create a label for each structure that will be segmented. For example, "hip" and "femur".
  • Go back to SegmentHumanBody module by clicking the green left arrow on the module toolbar. You are ready to segment now!

Automatic Segmentation

  • Click "Run Automatic Segmentation" button and wait until the processing is completed (it may take some time based on number of slices). SegmentAnyBone will run in automated mode and will segment bones in each slice it can detect. This is going to be a binary segmentation (e.g. bone vs no-bone). To assign labels to different object of interests in slices, you can use label assignment feature:
  • For each structure of interest: choose a label to assign (above the "Run automatic segmentation" button, for example: "hip") and click "Assign label (2D)" or "Assign label (3D)" button
    • 2D Label Assignment: Change label of the connected component only in the current slice. You should first place a prompt point on the object of interest whose label you want to change and click Assign Label (2D) button. Label of the object of interest will be changed only on that particular slice.
    • 3D Label Assignment: Change label of the connected component through consecutive slices in 3D. You should first place a prompt point on the object of interest whose label you want to change. Then, you should click Assign Label (3D) button. Label of the connected component through consecutive slices will be changed.

Prompt Based Segmentation

  • Select the label you want to segment from the dropdown list (for example "femur" as shown in the image above).
  • Click "Start Segmentation for Current Slice" button.

If it is the first time to segment a slice of this file, you need to wait for SegmentHumanBody to produce some files that will be used for the segmentation. After SegmentHumanBody generated these files, you can start putting prompt points or prompt boxes on the current slice. You'll be able to see the segmentation mask on 3D Slicer. Please click "Stop Segmentation for Current Slice" whenever you finish your segmentation for the current slice.

If you are not satisfied with the segmentation mask produced by SegmentHumanBody, you can edit it as you wish using the "Segment Editor" module of 3D Slicer.

Future Work

More segmentation models that aims to segment various type of tissues will be added in the future.

Citation

If you find our work to be useful for your research, please cite our paper:

@misc{gu2024segmentanybone,
      title={SegmentAnyBone: A Universal Model that Segments Any Bone at Any Location on MRI},
      author={Hanxue Gu and Roy Colglazier and Haoyu Dong and Jikai Zhang and Yaqian Chen and Zafer Yildiz and Yuwen Chen and Lin Li and Jichen Yang and Jay Willhite and Alex M. Meyer and Brian Guo and Yashvi Atul Shah and Emily Luo and Shipra Rajput and Sally Kuehn and Clark Bulleit and Kevin A. Wu and Jisoo Lee and Brandon Ramirez and Darui Lu and Jay M. Levin and Maciej A. Mazurowski},
      year={2024},
      eprint={2401.12974},
      archivePrefix={arXiv},
      primaryClass={eess.IV}