NKI-AI/direct-cvpr2022-recurrentvarnet

DIRECT — CVPR 2022 RecurrentVarNet

Checkpoints from the DIRECT release related to Recurrent Variational Network (Yiasemis et al., CVPR 2022).

• Paper: arXiv:2111.09639
• Project: projects/cvpr2022_recurrentvarnet

Layout

calgary_campinas/<model>.{yaml,pt}
fastmri_axt1/<model>.{yaml,pt}

Directory	Dataset	Link
calgary_campinas/	Calgary-Campinas brain (multi-coil) — main experiments & ablations	sites.google.com/view/calgary-campinas-dataset
fastmri_axt1/	fastMRI brain AXT1 — paper appendix	fastmri.med.nyu.edu · fastmri.org/dataset

Challenge Poisson-disk masks: challenge site · Hub NKI-AI/direct-mri-masks.

Calgary-Campinas split used in the paper: 47 / 10 / 10 volumes (train / val / test) after discarding outer slices. Subsampling uses the challenge Poisson-disk masks. fastMRI AXT1 experiments use random Cartesian undersampling.

Calgary-Campinas

RecurrentVarNet (shared weights and ablations *_noRSI, *_noSER, *_T11) plus comparison baselines (rim, xpdnet, unet, varnet, …). Each inference YAML pins a single acceleration under inference.dataset.transforms.masking (default 5×; 10× commented) and sets crop_outer_slices: true.

hf download NKI-AI/direct-cvpr2022-recurrentvarnet --local-dir ./cvpr_rvn

direct predict ./predictions \
  --cfg ./cvpr_rvn/calgary_campinas/recurrentvarnet_shared_weights.yaml \
  --checkpoint ./cvpr_rvn/calgary_campinas/recurrentvarnet_shared_weights.pt \
  --data-root /path/to/calgary_campinas \
  --num-gpus 1

fastMRI AXT1 brain (paper appendix)

These checkpoints are from the appendix of Recurrent Variational Network (arXiv:2111.09639) (Yiasemis et al., CVPR 2022) — not the main Calgary-Campinas experiments. PDF: arxiv.org/pdf/2111.09639.

Models: recurrentvarnet, recurrentvarnet_t6_h128_nl2, lpd, rim, unet, varnet (recurrentvarnet_t6_h128_nl2 is the appendix ablation: T=6, 128 hidden channels, nl=2).

Inference YAMLs default to 4× random Cartesian undersampling (center_fractions: [0.08]). Commented 8× settings (accelerations: [8], center_fractions: [0.04]) are under inference.dataset.transforms.masking. Keep exactly one active acceleration list.

direct predict ./predictions \
  --cfg ./cvpr_rvn/fastmri_axt1/recurrentvarnet.yaml \
  --checkpoint ./cvpr_rvn/fastmri_axt1/recurrentvarnet.pt \
  --data-root /path/to/fastmri/brain/multicoil_val \
  --num-gpus 1

Install DIRECT

git clone https://github.com/NKI-AI/direct.git
cd direct
conda create --name direct python=3.12
conda activate direct
pip install meson-python meson ninja
pip install --no-build-isolation -e ".[dev]"

The first argument to direct predict is the prediction output directory.

Citation

If you use these models or DIRECT, please cite the DIRECT toolkit and the method paper(s) below.

DIRECT

@article{DIRECTTOOLKIT,
  title={DIRECT: Deep Image REConstruction Toolkit},
  author={Yiasemis, George and Moriakov, Nikita and Karkalousos, Dimitrios and Caan, Matthan and Teuwen, Jonas},
  journal={Journal of Open Source Software},
  volume={7},
  number={73},
  pages={4278},
  year={2022},
  doi={10.21105/joss.04278},
  url={https://doi.org/10.21105/joss.04278}
}

Method

@inproceedings{yiasemis2022recurrentvarnet,
  title={Recurrent Variational Network: A Deep Learning Inverse Problem Solver applied to the task of Accelerated {MRI} Reconstruction},
  author={Yiasemis, George and Sonke, Jan-Jakob and S{\'a}nchez, Clara I. and Teuwen, Jonas},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2022}
}