Reducing Metal Artifact using Iterative Reconstruction in Industrial CT

Abstract

This paper presents a combination of use of gamma rays and iterative Computed Tomography (CT) algorithm to reduce metal artifacts also known as streak artifacts in Industrial Computed Tomography (ICT) imaging. ICT imaging method generally makes use of the popular Filtered Back Projection (FBP) algorithm. In a typical ICT data scan this method is seen to create streak artifacts. One of the reasons for streak artifacts is polychromatic nature of X ray beam causing beam hardening effect while other reasons like radiation scatter and photon starvation are also responsible. Use of gamma rays may help in countering beam hardening but still we find streaks in CT image. In addition to use of gamma rays, metal artifacts are reduced by employing iterative reconstruction algorithms like Maximum Likelihood Expectation Maximization (MLEM). MLEM is chosen because it is faster and guarantees convergence. Here we have studied the effect of employing gamma rays and MLEM algorithm on metal artifact at different iterations.