This software is used to preprocess MRI images that have been downloaded with the MIM software. The two main objectives of this program are:
- Reorder folder (ReorderFolders.py). This program will read every MRI series for different patients in different dates and will move them to an sequential numbering folders which are easier to analyze. It also anonymize a section of the folder names, to remove patient information.
- Preprocess data (Preprocessing.py). This program changes the original format of the images (DICOM) into a more commonly used format in the Medical Imaging world (NRRD). Additionally, with this program we separate the desired contours into separate binary masks. With this program you can also perform important enhancements on the images and contours like: normalization, bias correction, resampling, cropping, etc.
conda instal python=3.5
conda install -c conda-forge dicom
conda install -c simpleitk simpleitk
while read requirement; do conda install --yes $requirement || pip install $requirement; done < requirementsT.txt
pip3 install --user numpy dicom pandas
In order to run the code, copy the file Example_MainConfig.py into MainConfig.py and update the desired parameters
with information from your computer and your dataset.
Our common pipeline is that the cases downloaded from MIM are first stored in a folder MIM_ORIGINAL then,
the output of the first program ReorderFolders.py is stored in MIM_ORGANIZED_NO_DATE and the
output of the second program Preprocessing.py is stored in Preproc folder.