Software designed to convert MRI diffusion-weighted sequences into FLAIR sequences
Python 3.8
First, clone the git directory :
git clone https://github.com/NeuroSainteAnne/synthFLAIR.git
cd synthFLAIR/
Then install required modules:
pip install -r requirements.txt
pip install jupyterlab
Open the jupyter notebook
jupyter notebook
Open the synthFLAIR-Dataprep in order to convert you NIFTI files in a Memmap format for model training
Open the synthFLAIR-Train notebook and follow instructions to train and save your model.
Alternatively, download the pre-trained weights available here.
Once the model is trained and saved, you can convert 4D NIFTI DWI volumes into synthFLAIR volumes with the synthFLAIR-Convert tool.
Once the model is trained and saved, you can use the Orthanc module to automatically convert DWI volumes into FLAIR using a lightweight DICOM server. Follow the intructions at Orthanc Module Readme