This readme file was generated on 2025-12-05 by Artyom Tsanda # GENERAL INFORMATION * Title of Dataset: Supplementary Data for the Paper: “Deep Learning for Restoring MPI System Matrices Using Simulated Training Data” ## Author/Principal Investigator Information Name: Artyom Tsanda ORCID: 0009-0009-7765-4604 Institution: Institute for Biomedical Imaging, Hamburg University of Technology, Hamburg, Germany Address: Lottestraße 55, 22529 Hamburg, Germany Email: artyom.tsanda@tuhh.de ## Author/Associate or Co-investigator Information Name: Sarah Reiss ORCID: 0009-0006-4015-1200 Institution: Institute for Biomedical Imaging, Hamburg University of Technology, Hamburg, Germany Address: Lottestraße 55, 22529 Hamburg, Germany Email: sarah.reiss@tuhh.de ## Author/Alternate Contact Information Name: Konrad Scheffler ORCID: 0000-0002-6842-9204 Institution: Institute for Biomedical Imaging, Hamburg University of Technology, Hamburg, Germany Address: Lottestraße 55, 22529 Hamburg, Germany Email: konrad.scheffler@tuhh.de ## Author/Alternate Contact Information Name: Marija Boberg ORCID: 0000-0003-3419-7481 Institution: Institute for Biomedical Imaging, Hamburg University of Technology, Hamburg, Germany Address: Lottestraße 55, 22529 Hamburg, Germany Email: marija.boberg@tuhh.de ## Author/Alternate Contact Information Name: Tobias Knopp ORCID: 0000-0002-1589-8517 Institution: Institute for Biomedical Imaging, Hamburg University of Technology, Hamburg, Germany Address: Lottestraße 55, 22529 Hamburg, Germany Email: tobias.knopp@tuhh.de * Date of data collection: 2025-03-01 -- 2025-12-01 * Geographic location of data collection: Hamburg, Germany * Information about funding sources that supported the collection of the data: This project is funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) - SFB 1615 - 503850735. # SHARING/ACCESS INFORMATION * Licenses/restrictions placed on the data: Creative Commons Attribution 4.0 International License (CC BY 4.0) * Recommended citation for this dataset: To clarify what constitutes appropriate credit for specific parts of this dataset, please include the corresponding citation(s) listed below when your work makes use of data with the indicated name(s). If you use multiple files or all files from the dataset, please cite all applicable publications. Name-Specific Attribution Requirements: 1. Files starting with "measurements__snake" If you use files with this prefix, please cite: @article{maas_eq_model, author = {Maass, Marco and Kluth, Tobias and Droigk, Christine and Albers, Hannes and Scheffler, Konrad and Mertins, Alfred and Knopp, Tobias}, title = {Equilibrium Model With Anisotropy for Model-Based Reconstruction in Magnetic Particle Imaging}, year = {2024}, volume = {10}, pages = {1588--1601}, journal = {IEEE Transactions on Computational Imaging}, publisher = {IEEE}, doi = {10.1109/TCI.2024.3490381} } 2. Files starting with "measurements__resolution" If you use files with this prefix, please cite: @article{knopp_openmpi, title = {OpenMPIData: An initiative for freely accessible magnetic particle imaging data}, author = {Knopp, Tobias and Szwargulski, Patryk and Griese, Florian and Gr{\"a}ser}, Matthias, journal = {Data in Brief}, volume = {28}, pages = {104971}, year = {2020}, issn = {2352-3409}, doi = {10.1016/j.dib.2019.104971}, } 3. Files starting with "measurements__spiral" If you use files with this prefix, please cite: @article{mohn_resotran, title = {Characterization of the clinically approved MRI tracer resotran for magnetic particle imaging in a comparison study}, author = {Mohn, Fabian and Scheffler, Konrad and Ackers, Justin and Weimer, Agnes and Wegner, Franz and Thieben, Florian and Ahlborg, Mandy and Vogel, Patrick and Graeser, Matthias and Knopp, Tobias}, journal = {Physics in Medicine \& Biology}, volume = {69}, number = {13}, pages = {135014}, publisher = {IOP Publishing}, doi = {10.1088/1361-6560/ad5828}, year = {2024}, } 4. Files starting with "measurements__rectangle" If you use files with this prefix, please cite: @ARTICLE{szwargulski_multipatch, author={Szwargulski, Patryk and M{\"o}ddel, Martin and Gdaniec, Nadine and Knopp, Tobias}, journal={IEEE Transactions on Medical Imaging}, title={Efficient Joint Image Reconstruction of Multi-Patch Data Reusing a Single System Matrix in Magnetic Particle Imaging}, year={2019}, volume={38}, number={4}, pages={932-944}, doi={10.1109/TMI.2018.2875829} } 5. Any other file If you use any file other than from 1-4, please cite: @article{tsanda_deep_2026, title = {Deep {Learning} for {Restoring} {MPI} {System} {Matrices} {Using} {Simulated} {Training} {Data}}, doi = {10.1088/1361-6560/ae6016}, journal = {Physics in Medicine & Biology}, author = {Tsanda, Artyom and Reiss, Sarah and Scheffler, Konrad and Boberg, Marija and Knopp, Tobias}, year = {2026}, } # DATA & FILE OVERVIEW ## File List: ``` . ├── flatten.sh ├── measurements__noise__2D__noise.npy ├── measurements__noise__3D__chunk_1.npy ├── measurements__noise__3D__chunk_2.npy ├── measurements__noise__3D__chunk_3.npy ├── measurements__noise__3D__chunk_4.npy ├── measurements__noise__3D__chunk_5.npy ├── measurements__rectangle__mask.npy ├── measurements__rectangle__sm_measured.mdf ├── measurements__rectangle__u.mdf ├── measurements__resolution__sm_measured.mdf ├── measurements__resolution__u.mdf ├── measurements__snake__sm_averaged.mdf ├── measurements__snake__sm_noisy.mdf ├── measurements__snake__u.mdf ├── measurements__spiral__sm_measured.mdf ├── measurements__spiral__u.mdf ├── models__accelerated-calibration__3D__smrnet__checkpoint.pt ├── models__accelerated-calibration__3D__smrnet__config.json ├── models__denoising__2D__dncnn__checkpoint.pt ├── models__denoising__2D__dncnn__config.json ├── models__denoising__2D__rdn__checkpoint.pt ├── models__denoising__2D__rdn__config.json ├── models__denoising__2D__swinir__checkpoint.pt ├── models__denoising__2D__swinir__config.json ├── models__denoising__3D__dncnn__checkpoint.pt ├── models__denoising__3D__dncnn__config.json ├── models__denoising__3D__rdn__checkpoint.pt ├── models__denoising__3D__rdn__config.json ├── models__inpainting__3D__pconvunet__checkpoint.pt ├── models__inpainting__3D__pconvunet__config.json ├── models__upsampling__x2__2D__smrnet__checkpoint.pt ├── models__upsampling__x2__2D__smrnet__config.json ├── models__upsampling__x4__2D__smrnet__checkpoint.pt ├── models__upsampling__x4__2D__smrnet__config.json └── simulation_parameters.json ``` Files starting from "measurements__" contain measured or simulated MPI system matrices and noise realizations. Files starting from "models__" contain trained deep learning models for different restoration tasks. "simulation_parameters.json" contains the parameters used for simulation of training data. "flatten.sh" is a helper script to bring back the original hierarchy of measurement files after downloading. # METHODOLOGICAL INFORMATION ## Description of methods used for collection/generation of data: The data were generated as described in the accompanying publication: Tsanda, A., Reiss, S., Scheffler, K., Boberg, M., Knopp, T., 2026. Deep learning for restoring MPI system matrices using simulated training data. Phys. Med. Biol. https://doi.org/10.1088/1361-6560/ae6016 ## Methods for processing the data: The data are either the result of direct measurements using a magnetic particle imaging scanner or artifacts generated through training deep learning models on simulated data. ## Instrument- or software-specific information needed to interpret the data: The data accompany the repository: https://github.com/ibiResearch/mpi-sm-restoration