Efficient acquisition for periodic dynamic CT

In this work, we study the theoretical conditions on a X-ray computed tomography (CT) acquisition protocol to get high temporal and spatial resolution reconstructions of the slices of an object, under the hypothesis that the object has a cyclic evolution. We first show how the rotation time setting influences the theoretically reachable temporal resolution and how this setting should be optimized. Then for the discrete acquisitions, we propose efficient sampling schemes based on the generalization of the sampling conditions in helical CT. Both approaches, the adjustment of rotation speed and sampling conditions, are then combined to propose optimal acquisition protocols in order to reduce the amount of measured data and improve the quality of the reconstruction of an image sequence. These results could be applied to cardiac imaging, when the periodicity hypothesis is acceptable during the short acquisition time.