scilus / ismrm_2015_tractography_challenge_scoring

Scripts used to recreate the results of the ISMRM 2015 Tractography Challenge
16 stars 9 forks source link

ISMRM 2015 Tractography Challenge Scoring system

This system contains the scripts and tools that can be used to recreate the results of the ISMRM 2015 Tractography Challenge and to evaluate new datasets.

The release used to produce results for the website and paper is archived on Zenodo DOI.

Configuration and installation

Make sure the "pip" version is recent enough. You can update it with

pip install -U pip

Then, install the needed dependencies using the requirements file. The team recommends using a virtual environment (with the virtual env wrapper), but it is not mandatory. Once setup, run

pip install -r requirements.txt

You can then clone or download the scoring system. Once cloned or downloaded, go inside the ismrm_2015_tractography_challenge_scoring directory, and run

python setup.py build_ext --inplace
python setup.py install

Once those steps are all done, the system is configured.

Fetching the Ground Truth Dataset

To be able to run the scoring system, a directory containing the ground truth dataset is needed. It can be downloaded from the Tractometer website.

Scoring a tractogram

Once the ground truth dataset is unarchived (for example, to the scoring_data directory, one needs to create a directory where all results will be saved. Let's call it results for now.

Then, an example call to the scoring system will be

./scripts/score_tractogram.py YOUR_TRACTOGRAM_FILE scoring_data/ results/

where YOUR_TRACTOGRAM_FILE is replaced with the path of the tractogram file that will be scored.

Additional flags use to control the saving behavior of the script are available. Call score_tractogram.py -h to get the list of such flags.