Motion Compensated CT Reconstruction of the Head

Spontaneous patient motion occurring in head imaging can lead to severe image artifacts. In this study, we introduce a simulation framework for two different types of head motion and investigate the severity of the resulting artifacts under the assumption of perfectly known rigid motion parameters during motion compensated reconstruction for circular head CT. Error measures such as the mean absolute error and the structural similarity index are evaluated as a function of the rotation angle and speed during the data acquisition. Targeting mean absolute errors of 5 HU in the brain tissue of clinical data sets, motion up to 1.3° (6°) nodding motion amplitude can be tolerated in the normal (motion compensated) reconstruction.