An AI-based framework for the visualization and analysis of massive amounts of 4D tomography data for end users of beamlines

The core concept of the AI4D4E software framework Voxie is to work with 4D CT data compressed by a factor of 10 in a computer’s main memory, so that image quality remains virtually unaffected and 1–2 terabytes of 4D data can be visualized on a PC with 128 GB or 256 GB of RAM at the user’s workstation. With this Voxie framework, end users of beam guidance systems can utilize AI-based methods to process the enormous amounts of data from their 4D CT measurements. This includes improving image quality through artifact reduction, as well as reducing the volume of data and making it accessible to end users to assist them in interpreting the results. The projekt is funded by BMFTR with Grant No. 05D2022.

The developed framework allows working with compressed 4D datasets in memory. Below, two videos made on a laptop show moving through the slices or the timesteps of a 4D dataset are shown in real time.

The project focuses on the topics of artefact reduction, segmentation and visualization of large 4D datasets. The resulting methods should be applicable to data generated by both photon and neutron sources.

The 4D in-situ/operando X-ray tomography of a commercial zinc-air battery shows the dissolution of the zinc particles during discharge.
Photo: Helmholtz-Zentrum Berlin (HZB), Tobias Arlt, Ingo Manke