Correction of motion artifacts in linogram and projection reconstruction MRI using geometry and consistency constraints

Motion results in various artifacts such as blurring and streaks in clinical imaging of subjects based on reconstruction from projections. We model subject motion-induced artifacts due to scaling, translational and rotational motion. A correction algorithm based on the Ludwig-Helgason consistency conditions is derived here. These conditions are satisfied whenever the projection data are consistent. We apply the algorithm to simulated data collected on linogram (LR) and projection reconstruction (PR) geometries, and to real PR geometry data, in magnetic resonance imaging (MRT). The results show that motion-induced in-plane, interview artifacts can be reduced with application of the algorithm. The algorithm is general enough to be applied to certain other cases arising in tomographic imaging. (C) 1996 American Association of Physicists in Medicine.