Evaluation of Two Sinogram Interpolation Methods for Metal Artefacts Reduction in Computed Tomography

摘要

In the presence of metal implant, it can induce streak artefacts appearance in computed tomography (CT) images that can degrade image quality and lead to misdiagnosis. A comparative study of two different sinogram interpolation methods for metal artefact reduction in CT images is presented. The metal part of the sinogram was determined and removed by thresholding technique. The missing part of the metal trace on the sinogram was interpolated by two interpolation methods which were cubic spline and Laplace interpolation. Both methods were applied to the selected CT images of different metal geometry and complexity. The effectiveness of both interpolation methods was compared. The results showed Laplace interpolation method produced the better results of gap filling as compared to cubic spline interpolation. But, it produces lower signal-to-noise ratio (SNR) values.