Truncation Correction using a 3D Filter for Cone-beam CT

Recently, a novel method for region of interest (ROI) reconstruction from truncated projections with neither the use of prior knowledge nor explicit extrapolation has been published, named Approximated Truncation Robust Algorithm for Computed Tomography (ATRACT). It was derived by analytically reformulating the standard Feldkamp-Davis-Kress (FDK) algorithm into a reconstruction scheme that is by construction less sensitive to lateral data truncation. In this paper, we present and investigate a variation of the ATRACT that is to apply ATRACT in 3D by decomposing the ramp filter into the 3D Laplace filter and a 3D residual filter. ROI reconstruction can be readily realized by performing these two successive filters on projection data stack at once and followed by standard backprojection. Real data evaluation shows that the new method at least performs as well as the native ATRACT in terms of truncation correction. However, for off-center reconstruction, the linear gradient artifact arose in native ATRACT is essentially reduced by the new method.