Metal Artifact Reduction Based on Beam Hardening Correction and Statistical Iterative Reconstruction for X-ray Computed Tomography

Metal artifact is a main cause to degrade CT image quality, but there is still no standard solution to this issue. The cause of introduction of metal artifacts is due to several physical effects, in which beam hardening and noise are two major factors. Accordingly, in this paper these two factors are alleviated by using beam hardening correction based on polynomial fitting and statistical iterative reconstruction based on Poisson log-likelihood approach. Unlike other metal artifact reduction (MAR) methods by using iterative image reconstruction from polychromatic projection dataset, the proposed method in this work does not require a priori knowledge about the X-ray spectrum and attenuations of the materials to be reconstructed. A conventional linear interpolation MAR algorithm and two MAR methods based on beam hardening correction are performed for comparison. Simulation results illustrate that the proposed method can suppress metal artifacts greatly and restore low contrast tissues well.