MAR-SART: Metal artifact reduction for simultaneous algebraic reconstruction technique

In X-ray computed tomography metallic objects, located in the scanned area, lead to streak artifacts in the reconstructed image data. The use of less absorbing materials (like titanium) for implants is often not an option, so the common solution is a modification of the existing 3D reconstruction algorithms. Usually these methods are based on complete scans of the affected area with many projection images. However, some applications like intraoperativ 3D imaging can implicate a limited scan angle and a reduced number of images to reduce the radiation exposure. In this article a new approach for metal artifact reduction using a simultaneous algebraic reconstruction technique is proposed. Metallic components in the projection images are detected and replaced with synthetic projection data t o improve the 3D reconstruction quality. First results from simulations with a computer generated phantom are presented.