Experimental 2D fan-beam X-ray CT dataset of a laser-cut Shepp–Logan-like phantom

Summary This submission contains a two-dimensional X-ray CT dataset of a physical Shepp-Logan-like phantom acquired at the FleX-ray Lab of the Centrum Wiskunde & Informatica (CWI), Amsterdam, the Netherlands. The dataset is intended as a real experimental benchmark for 2D tomographic reconstruction, inverse problems, sparse-view CT, and learned reconstruction algorithms. The raw projection data, dark-field image, flat-field images, scan scripts, and scanner metadata are provided in the native FleX-ray scanner format. This dataset was used in [Thomsen et al. 2026]. Sample Information The sample is a physical Shepp-Logan-like phantom manufactured from a 6 mm-thick transparent polymethyl methacrylate (PMMA) plate. Elliptical inclusions were laser-cut into the plate. Two ellipses were left as air-filled cavities, while the remaining inclusions were filled with transparent deep-pour epoxy resin. The phantom therefore contains three material classes: PMMA, epoxy resin, and air. Apparatus The dataset was acquired using the custom-built FleX-ray laboratory X-ray CT scanner located at CWI in Amsterdam. The FleX-ray scanner is a flexible cone-beam CT system with a microfocus X-ray source and a flat-panel detector. Full details of the apparatus can be found in [Coban 2020]. Data Acquisition The acquisition used smooth continuous rotation over 360 degrees. Although the scanner is a cone-beam CT system, the selected detector ROI contains a single binned detector row, so the data can be used as a fan-beam-equivalent 2D CT dataset. Number of projections: 901Projection image size: 478 x 1 pixelsSinogram size: 901 views x 478 detector channelsExposure time: 29.997 msBinning: 4 x 4Binned detector pixel size: 0.2992 mm Geometry Source-to-object distance (SOD): 234.919688 mmSource-to-detector distance (SDD): 400.000800 mmMagnification: 1.702713Center of rotation: 239.000316 detector pixelsHorizontal center (HC): 239.000161 detector pixelsVertical center (VC): 0.424366 detector pixels X-ray Source Tube voltage: 70 kV set value, 69.694680 kV actual valueTube current: 600 uA set value, 597.069597 uA actual valueFocus mode: high-power List of Contents All raw projection and reference images are provided as 16-bit TIFF files. The data folder contains: • scan_000000.tif - scan_000900.tif: raw projection images • di000000.tif: averaged dark-field image • io000000.tif: averaged flat-field image acquired before the scan • io000001.tif: averaged flat-field image acquired after the scan • scan settings.txt: human-readable scanner metadata; this file is used by the accompanying code to read the scan geometry • data settings XRE.txt: scanner metadata • Acquisition settings XRE.txt: acquisition log • data set settings.ini: a snapshot text file of basic geometry information at the start of a scan • script.txt: scan control script • script_executed.txt: executed scan commands Note on scanner metadata files The FleX-ray scanner produces several metadata, configuration, and log files during acquisition. In the accompanying preprocessing and reconstruction code, the scan geometry is read from scan settings.txt, and the geometry values reported in this Zenodo record are based on that file. The other metadata, configuration, and log files are included as part of the complete native FleX-ray acquisition output. They are provided to document the acquisition provenance and to support alternative processing pipelines, which may parse different FleX-ray/XRE metadata files. Some information is duplicated across these files and may be stored in different formats. Additional Links These data were acquired at the FleX-ray Lab of the Computational Imaging group at Centrum Wiskunde & Informatica (CI-CWI). For example usage, preprocessing, and reconstruction code, see the code repository accompanying [Thomsen et al. 2026]: https://github.com/ezgidetu/DDS-experimentalCT The corresponding paper is available at: https://arxiv.org/abs/2602.12755 Contact Details For more information or guidance in using this dataset, please contact: • [e-mail] • [e-mail] References [Coban 2020] S. B. Coban, F. Lucka, W. J. Palenstijn, D. Van Loo, and K. J. Batenburg, “Explorative imaging and its implementation at the FleX-ray Laboratory,” Journal of Imaging, 6(4), 18, 2020. https://doi.org/10.3390/jimaging6040018 [Thomsen et al. 2026] N. J. Thomsen, X. Wang, F. Lucka, and E. Demircan-Tureyen, “Towards reconstructing experimental sparse-view X-ray CT data with diffusion models,” arXiv:2602.12755, 2026. https://arxiv.org/abs/2602.12755