Rigid motion correction for head CT imaging

During a CT scan, the patient should not move since the algorithms used to compute the three dimensional images assume that the measured data provide information about a motionless object. However, for some patients it can be very difficult to stay motionless for several seconds or longer. In brain scanning, this problem is most prominent in psychiatric patients and children. Consequently, motion artifacts are often observed in these patients. These artifacts can be severe enough to make the images useless for diagnosis, and therefore patients are sometimes sedated or anaesthetized to prevent motion artifacts.

The aim of this project is to develop a clinically useable motion correction procedure for CT brain imaging. To this end, various approaches will be investigated: data-driven motion estimation isonly one example. Particular emphasis will be given to the accuracy of the developed motion-correction method. In addition, aiming at a clinically usable technique, methods to reduce the associated computational burden will be evaluated.