Development of the Projection-Based Material Decomposition Algorithm for Multienergy CT

Current advances in the development of hybrid pixel detectors allow analyzing an energy spectrum of the incoming X-ray radiation with a keV resolution. This spectral information can be utilized to determine the material composition of the studied object based on the known material attenuation dependency on energy. The algorithm discussed and implemented in this article solves an optimization problem with a cost function based on Beer's Law. A successful application of this procedure to the real data requires an accurate model of the detector response. A Monte-Carlo simulation of the registration process in the Timepix3-based detector is performed for the generation of the detector signal corresponding to a monochromatic beam depending on the detector properties. The material decomposition algorithm is applied to simulated data with a response function similar to the real detector. The obtained results are in accordance with the expected values of material concentrations, and the variance mostly depends on a statistical error of the simulation and properties of the detector. An application of the implemented algorithm to experimental data is attempted but the results contain qualitative and quantitative errors. Shortcomings of the current implementation and possible improvements of the detector model and decomposition procedure are discussed.