Skip to content

Latest commit

 

History

History
106 lines (79 loc) · 6.71 KB

README.md

File metadata and controls

106 lines (79 loc) · 6.71 KB
Raw

SlicerCART

SlicerCART: Configurable Annotation and Review Tool.

Description

  • 3D Slicer extension
  • Adapted from code of Dr. Laurent Létourneau-Guillon and his team in ICH_SEGMENTER_V2, SlicerCART, and Brain_Extraction. This is an effort to create a unified code for a configurable 3D Slicer extension.
  • Inspired from Neuropoly's workflow.
  • This tool is made to improve manual segmentation and classification workflows across different teams.

Keywords: medical imaging, manual segmentation, manual correction, workflow, ground-truth segmentation, quality control

Abbreviations:

  • MRI - Magnetic Resonance Imaging
  • CT - Computed Tomography
  • BIDS - Brain Imaging Data Structure
  • GUI - Graphical User Interface
  • QC - Quality Control

Rationale:

  • Manual segmentation and classification tasks are required in the research setting related to medical imaging artificial intelligence tool development
  • An open-source solution for such tasks would better benefit the research setting
  • Actual open-source solutions that enable imaging viewing and annotation are not optimal from an end-user standpoint (especially from various background), increasing the already high burden of manual segmentation and classification tasks
  • A workflow aimed to efficiently navigate through a dataset while performing manual segmentation / correction, including revision steps and robust annotation consistency assessment is crucial for handling large amount of data and provide the best ground-truth references segmentation as possible.

Module specific functions (in details)

This module has been adapted to perform several tasks. Among other things, it allows the user to:

  • Customize configuration preferences (see below for details)
  • Customize keyboard shortcuts
  • Identify the name, degree and revision step related to the human annotator
  • Select folder of interest where volumes are saved (and possibly impose BIDS)
  • Select the output folder where processing and work is preferred to be saved
  • Select a ground-truth folder where references studies can be used for iterative self-assessment
  • Display in the GUI a case list of all the studies of interests for the segmentation task (*from a site directory or a customized list)
  • Select from the GUI case list any volume of interest to display
  • Navigate through case list from next and previous buttons
  • Load automatically the first remaining case for segmentation in a customized list
  • Create automatically all required segments that may be used according to the project configuration each time a volume is displayed
  • Toggle interpolation of the volume loaded
  • Execute multiple automated functions when saving segmentation masks for a given volume. Indeed, the automated functions:
    • Save segmentation masks in the selected output folder with volume file hierarchy
    • Track the different versions (save the following version if previous version(s) already exist(s)) *N.B. limitation to 99 versions for a single volume
    • Save a .csv file with segmentation statistics (e.g. subject, annotator's name and degree, revision step, date and time, total duration, duration of each label annotation)
    • Save a .csv file with classification statistics (e.g. subject, annotator's name and degree, revision step, date and time, checkboxes / dropdown / free text fields)
    • Go to the next remaining case and make it ready to segment without any further action
  • Load a pre-existing segmentation for further modification (will be saved as a new version of the segmentation)

Requirements

  • MacOS Sonoma or Sequoia is recommended
  • A working version of 3D Slicer.
    • N.B. The version currently used to develop this module is the version 5.6.2 (most stable release as of 2024-10-22). The version 5.2.2 has also been used.

This module has been developed on:

  • MacOS Sonoma version 14.1.1 and Sequoia 15.0.1
  • 3D Slicer version 5.6.2

Although it may work on other versions and/or operating system, please note that it has not been tested.

Installation steps

If previous version of Slicer ---» delete the app. If on MacOS, you can do it by doing right lick on the app in the application folder --- move to trash --- go to trash --- empty the trash --- restart your computer.

  1. Install 3D Slicer.
  2. Clone this repository in the location of your choice.
  3. Then, go in the finder --- find the file SlicerCART.py file, and copy the pathname.
  4. Open 3D Slicer.
  5. Activate the checkbox Enable developer mode in Edit -> Application Settings -> Developer -> Enable developer mode.
  6. Add the path of the SlicerCART.py file in Edit -> Application Settings -> Modules -> Additional module paths.(N.B. 1) You must have the file: if it is the folder path, then the module will not work; 2) in the Additional modul path section, the path copied might be shown to the folder: this is a Slicer thing, and should not affect working property of the module if it was the file that you copied).
  7. The module can be found under Examples -> SlicerCART: the module should now be opened (N.B. 1) If first use, you may have additional requirements to install. A pop-up window from Slicer advertising you should pop-up if so: just click ok).
  8. (Optional) Set the SlicerCART module to launch at 3DSlicer startup. To do so, go to Edit -> Application Settings -> Modules -> Default startup module
  9. There might be errors. These would be seen in the Python Console: if any errors, we highly recommend you to fix them before any further use!

Trouble shooting

  • Qt might need to be installed. The first five steps of the following procedure might be useful for this: procedure.

Other extensions that could be useful

  • SlicerJupyter to be able to use Jupyter Notebooks connected to 3D Slicer.

Documentation

TODO (after sufficient development has been made)

Video tutorials

TODO (after sufficient development has been made)

Other resources

Contributors

Contributors, including:

  • Laurent Létourneau-Guillon
  • Emmanuel Montagnon
  • An Ni Wu
  • Maxime Bouthillier
  • Delphine Pilon
  • Neuropoly Team