Metal artifact reduction with DCT-domain gap-filling method

The highly attenuating materials in the field-of-view result in high sinogram bin values which cause the so called metal artifacts in the reconstructed images. These artifacts degrade the quality of the images especially in X-Ray computerized tomography and accuracy of the attenuation correction factors necessary for quantitative positron emission tomography. In this study, we investigate the capability of our previously published DCT-domain gap-filling method for metal artifact reduction. We tested the method with the numerical jaw phantom in which the metal dental fillings were modeled with high intensity values. The sinogram bins corresponding to the line integrals calculated through the metal implants were assumed as the missing sinogram bins. The DCT-domain coefficients corresponding to these sinogram bins were then masked out with the dedicated DCT-domain filter and estimated in the iterative procedure of the gap-filling method. With the DCT-domain gap-filling method, we observed significant visual reduction of metal artifacts in the reconstructed images. These promising preliminary results presented in this study show the natural extension of the DCT-domain gap-filling method for the metal artifact reduction.