With this container you can start to reconstruct MR data straight away - it has Gadgetron, BART, MRIReco.jl and SIRF and all their dependencies installed.
In order to build the docker image, clone this repository
git clone https://github.com/ckolbPTB/OpenSourceMrRecon.git
Change into the directory
cd OpenSourceMrRecon
and build docker image
docker build -t os_recon .
Here we are tagging the image with os_recon. Feel free to use any other tag, but make sure you select the correct image when starting the container.
The build process will take some time (on a normal laptop around 1.5 hours) and the final image will be around 9GB in size.
Once the build is successfully finished, start the container in interactive mode
docker run -it os_recon
Enable the conda environment with all dependencies installed
source /opt/conda/bin/activate RecoEnv
Now you are ready to go!
If you want to reproduce the figure showing the different reconstructions of the cardiac spin echo data, then you only need to mount a folder when starting the container where the figure can be saved.
docker run -it -v full_path_to_your_folder:/recon_output os_recon
Once the container has started, make sure to enable the conda environment with all dependencies installed
source /opt/conda/bin/activate RecoEnv
Start a gadgetron server (and hit enter afterwards to return to the command line)
gadgetron &
Run the python script with all reconstructions
python example_code/run_open_source_recon.py
You will see lots of output in the terminal but that's all fine. Once the script has finished you should find the figure open_source_recon.png in full_path_to_your_folder.