Motion Tracking by Constrained Reconstruction of Orthogonal Projections
Huang Y, Wright G
University of Toronto, Sunnybrook & Women's College Health Sciences Centre
The AUCTION algorithm (Assignment and Update with CorrelaTION) was developed to reconstruct 3D data sets from orthogonal projections for time-resolved MR angiography. Correlations with a 3D spatial constraint are used to solve the ambiguity problems of the reconstruction. In the presence of motion, however, the data set forming the spatial constraint will be blurred. The mis-registration of the 3D spatial constraint to a particular pair of orthogonal projections may introduce ambiguities in the AUCTION reconstruction. In this work, the motion correction method for the AUCTION algorithm is investigated. The effectiveness of motion tracking by this technique is demonstrated in phantom studies.