Here we pursue the development of advanced reconstruction and visualisation methods for brilliant X-ray diagnostics based on the systems of Projects C.1.3 and C.1.4. The ability to acquire multi-modal X-ray imaging data sets containing attenuation, phase-contrast and dark-field information all at once necessitates new suitable reconstruction methods using and exploiting the added dimensions of the data. Directional dark-field tensor imaging in particular requires new approaches to the reconstruction problem. The reconstruction methods will be based on the iterative series expansion approach.
In a first step, a mathematical and physical model suitable for the joint attenuation, phase-contrast and dark-field acquisitions as well as the directional dark-field tensor data will be formulated. In a second step, iterative solvers for the resulting system equation will be adapted to the problem at hand in order to ensure convergence and acceptable convergence rates, in particular in view of initialisation using single component reconstructions. Systematic evaluations of the reconstruction methods will be performed on suitable phantoms and samples in cooperation with Projects C.1.3 and C.1.4 as well as on pre-clinical data from Research Areas C.2 and C.3, as they become available. The reconstruction framework will be available for use by the other relevant projects.
The results of the reconstructions include information about multiple modes (attenuation, phase contrast and dark-field) or in the case of directional dark-field imaging an anisotropy tensor. This is quite a lot more information than medical doctors are used to from regular X-ray CT necessitating the development of suitable visualisation methods for this new kind of data. In a close feedback loop with physicians from Research Areas C.2 and C.3 the final goal of this project is to devise and develop a visualisation tool for multi-modal X-ray imaging applications facilitating clinical translation.