Evaluation of Spectrum Mismatching Using Spectrum Binning for Statistical Polychromatic Reconstruction in CT

In CT, the nonlinear attenuation characteristics of polychro- matic X-rays cause beam hardening artifacts in the reconstructed im- ages. Statistical algorithms can effectively correct beam hardening ar- tifacts while providing the benefit of noise reduction. In practice, a big challenge for CT is the difficulty at acquiring accurate energy spectrum information, which hinders the efficiency of beam hardening correction approaches that require the spectrum as prior knowledge such as the statistical methods. In this paper, we used proposed energy spectrum binning approach for reducing prior knowledge from full spectrum to three energy bins to compare the results when applying parameters op- timized for one spectrum to data measured using a different spectrum.