Axial Cone Beam Reconstruction with Asymptotic Trajectory Extrapolation Implemented by 3-D Weighting Scheme

Abstract

Axial scan along circular trajectory has been the most desirable in clinical and preclinical applications. However, due to the violation of the data sufficiency condition, the reconstruction accuracy of axial scan deteriorates rapidly with increasing detector z-dimension. A number of algorithms have been proposed thus far to improve the reconstruction accuracy. It is well understood that, the larger the radius of circular trajectory, the more accurate the reconstruction. Ultimately, the reconstruction becomes accurate if the trajectory radius reaches infinite. Inspired by such observations, an asymptotic weighting scheme is proposed here for axial scan at large cone angles to improve reconstruction accuracy. In practice, there may be other ways to implement the asymptotic 3-D weighting scheme, and we present one approach in this paper with preliminary results. Further investigation may enable this asymptotic weighting scheme to improve reconstruction accuracy significantly, enabling numerous applications in which axial scan at large z-dimension is desired.