ROI--Driven CT Trajectories

Abstract

The CT field of measurement (FOM) in the transversal direction is typically determined by the fan angle and by the distance of the object to the focus. Large objects that do not fit within the FOM can be assessed only partially (and one has to cope with incomplete Radon data due to data truncation). We propose an algorithm to find a scan trajectory that allows to assess arbitrarily shaped and arbitrarily sized ROIs that may correspond to the object support -- and therefore the method solves the problem of scanning objects larger than the FOM in an elegant and efficient way -- or to any user--defined ROI within the object. The study was inspired by micro--CT imaging where motorized x--y--translation stages allow us to continuously shift the object during the scan as a function of the rotation angle (figure 1). The approach includes an automatic exposure control (AEC) technique that takes care of redundancy weighting and that either performs optimally in terms of object dose (in--vivo) or in terms of detector dose (in--vitro).