Regularised FISTA reconstruction algorithm for X-ray CT (FISTA-tomo)

Regularised FISTA-type iterative reconstruction algorithm for X-ray tomographic reconstruction with highly inaccurate measurements.

Software highlights:

• Tomographic projection data are simulated without the "inverse crime" using TomoPhantom. Noise and artifacts (zingers, rings) can be modelled and added to data.
• Simulated data reconstructed iteratively using FISTA-type algorithm with multiple "plug-and-play" regularisers from CCPi-RegularisationToolkit
• Presented FISTA algorithm offers novel modifications: convergence acceleration with ordered-subsets method, PWLS, Group-Huber and Students't data fidelities to deal with noise and image artifacts
• Various projection (2D/3D) geometries are supported and real data provided to demonstrate the effectiveness of the method