Multi-Material Beam Hardening Correction(MMBHC) in Computed Tomography

In computed tomography (CT), the nonlinear attenuation characteristics of polychromatic X-rays cause beam hardening artifacts in the reconstructed images. State-of-theart methods to correct the beam hardening effect are mostly single material precorrections (e.g. water-precorrection), which are far less efficient when more than one material is present in the field of measurement. The use of those techniques is limited by specific restrictions to the objects, computational loads, and inaccurate segmentations. In this paper, we present a practical multi-material beam hardening correction(MMBHC) approach that employs material decomposition technique maintaining CT values from dual-energy CT. This separates single energy CT images into spatial density images and images containing material information. The segmentation maintains the original X-ray attenuation coefficients, such that the original CT attenuation image can be exactly recovered. Therefore, segmentation errors, which result in invalid material properties to the voxel, only have minor effects on the beam hardening correction and do not cause an atypical image impression or introduce additional artifacts. A multi-material beam hardening correction procedure is formulated to iteratively correct the artifacts but shows satisfactory image quality after the first iteration. Based on experiments with simulated CT data, it is shown that the proposed method can efficiently reduce beam hardening artifacts. In addition to the performance benefits, our approach can be flexibly applied to imaging geometries and achieve efficient, fully 3D reconstructions.