Method for system-independent material characterization from spectral X-ray CT

摘要

We propose a method for material characterization using Spectral X-ray Computed Tomography (SCT). Our SCT method takes advantage of recently-developed MultiX ME 100 photon counting detectors to simultaneously measure the energy dependence of a material's linear attenuation coefficient (LAC). Relative electron density (rho(e)) and effective atomic number (Z(e)) are estimated directly from the energy-dependent LAC measurements. The method employs a spectral correction algorithm and automated selection and weighting of the energy bins for optimized performance. When examining materials with Z(e) <= 23, this method achieves accuracy comparable to traditional dual-energy CT, which is often realized through consecutive data acquisitions, and is compatible with any spectral detector. The method disregards data in photon starved energy channels improving the detection of highly attenuating materials, compared to techniques that use energy integrating detectors.