MONAI in DXJupyterLab

Using MONAI Core, MONAI Label/3D Slicer (SlicerJupyter) via DXJupyterLab

Medical Open Network for AI(MONAI) is a framework built for deep learning in healthcare imaging. To use MONAI on the DNAnexus Platform, run DXJupyterLab with the MONAI_ML feature, which includes:

• MONAI Core: PyTorch-based framework for deep learning in healthcare imaging.

• MONAI Label: An intelligent image labeling and learning tool designed to create training datasets and build AI annotation models. It provides a server-client framework that integrates with imaging viewers.

• 3D Slicer: An open-source software designed for the visualization, processing, and analysis of medical, biomedical, and other 3D images. In a Jupyter environment, 3D Slicer is accessible through the SlicerJupyter kernel and acts as a client for the MONAI Label server.

The MONAI Core, MONAI Label, and 3D Slicer (SlicerJupyter) come pre-installed with the DXJupyterLab MONAI_ML feature option.

For the full list of pre-installed packages, see the DNAnexus JupyterLab in-product documentation.

Using MONAI Core

For sample Jupyter notebooks and tutorials, see the official project MONAI tutorials.

You can find technical documentation for MONAI Core.

Using MONAI Label with 3D Slicer

For examples showing how to use 3D Slicer with MONAI Label, see the following sample Jupyet notebooks in DNAnexus OpenBio repository:

• Radiology Auto-Segmentation and Training with MONAI Label and 3D Slicer (NIfTI/CT): Demonstrates auto-segmentation and model training on NIfTI CT spleen data using MONAI Label and 3D Slicer (SlicerJupyter).

• Whole Brain Segmentation with MONAI Label and 3D Slicer (DICOM/MRI): Shows auto-segmentation and model training on DICOM MRI brain data, including DICOM-to-NIfTI conversion and interactive annotation in 3D Slicer.

For general examples and tutorials on using MONAI Label and 3D Slicer (SlicerJupyter), explore the following GitHub repositories:

• MONAI Label tutorials: Project-MONAI/tutorials/monailabel

• 3D Slicer (SlicerJupyter) example notebooks: Slicer/SlicerNotebooks