Intapatient, rigid+ affine+ b-spline transformation, mutual information
The registration was used to align organs in the abdomen of sequential cine-MR dynamics to the reference dynamic, randomly chosen from the cine after entering the steady-state. This registration was used to obtain deformable intrafraction motion from 2D cine-MR images as a potential deep learning ground-truth to learn deformtion vector fields.
For parameter files see the Elastix Model Zoo repository on GitHub.
elastix version: 4.800
Command line call:
elastix -f -m -p param_rigid.txt -p param_affine.txt -p param_bspline.txt -out
where is the path to fixed image, the reference cine MRI dynamic. The is the path to a cine-MR dynamic of the same imaging session. is the output directory.
Terpstra ML, Maspero M, D'Agata F, Stemkens B, Intven MP, Lagendijk JJ, Van den Berg CA, Tijssen RH. Deep learning-based image reconstruction and motion estimation from undersampled radial k-space for real-time MRI-guided radiotherapy. Physics in Medicine & Biology. 2020 July 30; 66(15). doi:https://doi.org/10.1088/1361-6560/ab9358.
© 2020 Viktor van der Valk