X-ray CT images of Lithium batteries for data augmentation in ML-based segmentation

This dataset was generated using the Python script available in the repository https://github.com/TomographicImaging/gVXR-training-dXCT2026. It was written for the workshop day at the dimensional X-ray Computed Tomography (dXCT), on the 16th of June 2026. This hands-on workshop provides an end-to-end workflow for X-ray Computed Tomography (XCT), combining physics-based simulation with machine-learning segmentation. XCT projection data were simulated using the gVXR package and reconstructed using CIL. It makes use of the "XT H 225" twin and loops over: - 6 samples of lithium batteries, - SOD: 150 mm +/- 20%, - Current: 160 uA +/- 30%, - Exposures: [0.5, 1.0, 1.42, 2.0] seconds, and - Voltages in the range [180, 225]. The dataset is used used for data augmentation to train machine-learning segmentation models.