RootPainter3D

This software is not approved for clinical use. Also please See LICENSE file.

Described in the paper "RootPainter3D: Interactive-machine-learning enables rapid and accurate contouring for radiotherapy".

Preprint is available at: https://arxiv.org/pdf/2106.11942.pdf

Published version is available at: http://doi.org/10.1002/mp.15353

Server setup

For the next steps I assume you have a dedicated GPU with CUDA and cuDNN installed. A recent version of python 3 is required.

1. Clone the RootPainter3D code from the repository and then cd into the trainer directory (the server component).

git clone https://github.com/Abe404/RootPainter3D.git
cd RootPainter3D/trainer

2. To avoid alterating global packages, I suggest using a virtual environment:

python -m venv env --clear

Note: Make sure to use python3. You may need to write python3 instead of python to do this.

3. Then activate the virtual environment:

On Linux:

source ./env/bin/activate

On Windows:

env\Scripts\activate.bat

4. Install PyTorch by following the instructions at the pytorch website

5. Install the other dependencies in the virtual environment:

pip install -r requirements.txt

6. Then, simply run RootPainter by:

python main.py

This will first create the sync directory.

You will be prompted to input a location for the sync directory. This is the folder where files, including instructions, are shared between the client and server. I will use ~/root_painter_sync

The RootPainter3D server will then create folders inside ~/root_painter_sync and start watching for instructions from the client.

You should now be able to see the folders created by RootPainter (datasets, instructions and projects) inside ~/root_painter_sync on your local machine.

Client setup

1. Clone the RootPainter3D code from the repository and then cd into the painter directory (the client component).

git clone https://github.com/Abe404/RootPainter3D.git
cd RootPainter3D/painter

2. To avoid alterating global packages. I suggest using a virtual environment. Create a virtual environment

python -m venv env

And then activate it.

On linux/mac:

source ./env/bin/activate

On windows:

env\Scripts\activate.bat

3. Install the other dependencies in the virtual environment

pip install -r requirements.txt

4. Run the client.

python src/main.py

To interactively train a model and annotate images, you will need to add a set of compressed NIfTI images (.nii.gz) to a folder in the datasets folder and then create a project using the client that references this dataset.

For more details, see the published article.

Contribution Guidelines

• Contibutions are welcome
• It's required to read the two published RootPainter papers first, to understand what the system does and how it works. I suggest starting with https://nph.onlinelibrary.wiley.com/doi/full/10.1111/nph.18387 (preprint at https://www.biorxiv.org/content/10.1101/2020.04.16.044461v1.full) and then moving onto https://arxiv.org/abs/2106.11942
• Contributions should be made via pull requests
• The GPL header should be clearly present at the top of any new files. For an example see: https://github.com/Abe404/RootPainter3D/blob/master/trainer/metrics.py
• The copyright header should include the current year and yourself or 'Anonymous Contributor' if you wish to remain somewhat anonymous to future readers of the code.

When we have tests, they should pass.