Craniosynostosis Distance Maps (CD-Map)

Images taken from the paper

Overview

This is a Python toolbox to create distance maps, e.g. for a CNN-based classification of head deformities.

We recommend using a virtual environment to manage package dependencies. This repository can be combined with the instances of the statistical shape model available at Zenodo.

Requirements and installation

The toolbox only depends on Python and should therefore be cross-platform.
Only some dependencies need to be installed, preferably in a virtual environment.

Setting up the virtual environment

The distance maps toolbox requires python3. We recommend using a virtual environment to install the additional dependencies. Since Python 3.5, the use of venv is recommended: For setting up the environment on any platform, we refer to the official documentation. Alternatives to the native package management include for example conda.

On Ubuntu, you can use for example:

python3 -m venv $HOME/venv/cdmap

To activate the virtual environment, use:

source $HOME/venv/cdmap/bin/activate

Installing the python packages

For installing the python packages we recommend pip which usually ships with Python by default. Install the following dependencies inside your virtual environment:

pip install numpy scikit-image vtk

Alternatively, we provide a requirements file which can also be installed using pip. However, this will install more packages than required.

We used Python version 3.7, but any reasonably recent version of Python3 should be fine.

Demo

Inside your virtual environment, run:

python3 demo.py

This will create the distance maps for the four mean shapes from the statistical shape model and should be enough to get you going.

License

All source code is subject to the terms of the General Public License 3.0.

References

If you use our code, please cite the corresponding paper: (https://ieeexplore.ieee.org/document/10129889) pdf