Cone-beam Gel phantom data for dynamic X-ray tomography

The gel phantom was constructed to simulate the flow of iodine-based contrast agents in plant stems, whilst being more tolerant to the high radiation dose of X-ray tomography scanning. A thorough explanation of the phantom, its motivation and the measurement setup can be found in arXiv:2003.02841. Attached is a brief documentation which focuses primarily on the higher dimensional 3D files and the key differences between the two data sets.

The primary measurements consist of 17 consecutive time frames, with an initial state of no contrast agent followed by a steady increase thereof, primarily via downward diffusion into the gel body over time. Each time frame consists of 360 projections acquired with a cone-beam micro-CT scanner, resulting in a sequence of 3D cone-beam sinograms. Moreover, the first time frame is provided with higher angular sampling of 720 projections, and an additional 18th time frame with 1600 projections.

The data set contains the following MATLAB files:

• 3dGelPhantomData_b4.mat
• 3dGelPhantomData_b8.mat
• 3dGelPhantomData_b16.mat
• 3dGelPhantomExtraFrames_b4.mat
• 3dGelPhantomExtraFrames_b8.mat
• 3dGelPhantomExtraFrames_b16.mat

And a more detailed documentation: ConeBeamGelPhantomDocumentation.pdf (v1.0)

The proper use of these files requires the ASTRA Toolbox, the Spot Linear-Operator Toolbox, and the HelTomo Toolbox. In addition the operators and reconstruction algorithms provided by ASTRA for 3D geometry require CUDA capable graphics card (GPU) due to large amounts of (parallel) computations. The older 2D data set provides a computationally lighter and more user-friendly starting point for inexperienced users.