Spectral Signature and Correction of Scattered Radiation in Energy-Resolved X-ray Imaging

X-ray imaging is a powerful tool for material identification and characterization with applications in a wide range of fields including medicine, biology, geoscience, and security. Since the discovery of x-rays in 1895, gradual improvements to imaging technology have lead to some major milestones such as computed tomography (CT). More recently, pioneering advancements in x-ray detector technology have made possible photon counting detectors (PCDs) with energy-resolving capabilities. Exploiting spectral information of the incident radiation promises revolutionary approaches to material identification and characterization. However, radiation that scatters from the object and reaches the detector is a long-standing problem that reduces image quality and quantitative accuracy. Previous studies to characterize and account for the scattered radiation have been limited to conventional x-ray imaging with energy-integrating detectors (EIDs).

The purpose of this research is two-fold: i) determine the spectral characteristics of the scattered radiation and the impact on quantitative spectral imaging and ii) develop an energy-sensitive scatter correction method to compensate for these inaccuracies. Through Monte Carlo simulation and experimental validation, the spectral characteristics of scatter are evaluated for a large scope of imaging parameters including: the object geometry and composition, object-to-detector distance, x-ray source distribution, and detector type. The impact of the scattered radiation was evaluated by estimating the energy-dependent attenuation properties of clinically-relevant materials. When left uncorrected, scattered radiation results in severe quantitative inaccuracies which can limit proper material identification.

The next objective was applying these characteristics to develop an energy-sensitive scatter correction that compensates for the inaccuracies due to scatter. Our method derives from the physical understanding of scatter interactions to estimate the spectrally-dependent scatter maps. This method was applied in the context of contrast-enhanced mammography, which showed accurate quantitative restoration of iodine targets in breast-like phantoms. This particular scatter correction technique is appealing as it does not require any modifications to the acquisition process or beam path. The versatility of the energy-sensitive scatter estimation technique also suggests further utility in other x-ray imaging applications such as tomosynthesis and computed tomography.