This repository provides a reference implementation of a binning reconstruction for 3D stack-of-spiral imaging using the Gadgetron's opensource framework. This implementation is presented in Self-gated 3D stack-of-spirals UTE pulmonary imaging at 0.55T.
A reference dataset is provided within the docker image and contains acquisitions, gradient waveforms, and header stored in the MRD format.
The MRD app can be downloaded from Docker Hub at https://hub.docker.com/r/gadgetronnhlbi/stack-of-spiral-mrd-app. In a command prompt on a system with Docker installed, download the Docker image:
docker pull gadgetronnhlbi/stack-of-spiral-mrd-app
Start the docker container on a system with gpus [ This implementation of binned stack of spiral reconstruction requires cuda ]
docker run --gpus all --name=mrd-app-sos -ti -p 9002:9002 --volume=/home/$USER/mrdata/:/opt/data --restart unless-stopped --detach gadgetronnhlbi/stack-of-spiral-mrd-app
Run the reconstruction using the provided test data within the container. First connect to the running container:
docker exec -ti mrd-app-sos bash
Then run the reconstruction with in the container
gadgetron_ismrmrd_client -p 9002 -f test_data/test_data_stack_of_spiral.h5 -c spiral_inline_binning.xml -o /opt/data/reconstructed_images.h5
You can monitor the logs in a seperate terminal by running
docker logs -f mrd-app-sos
The reconstructed images are written in /home/$USER/mrdata/. These can be viewed using a h5 viewer. There are two series of reconstructed images 1) fully sampled images 2) 40% stable binned data.