Two-step metal artifact reduction using 2D-NFFT and spherically symmetric basis functions

In computed tomography metal objects lead to inconsistencies in the acquired data which result in image artifacts after reconstruction. To enhance the image quality and allow for a more accurate diagnosis, a metal artifact reduction is required. Approaches may base on a recomputation of metal influenced values or use an alternative image reconstruction technique, which is less sensitive to inconsistent raw data than the standard filtered backprojection. Here, a two-step algorithm for metal artifact reduction is proposed, which combines a 2D NFFT-based interpolation and an MLEM reconstruction with spherically symmetric basis functions (blobs). Reconstructed images have been evaluated visually and quantitatively. Experiments show that combination of a 2D NFFT-based interpolation and blobs offers an effective strategy for enhanced metal artifacts suppression.