Hari Om Aggrawal: X-ray computed tomography is a mature technology that is used in many applications where cross-sectional images of high resolution provide useful information.
Medical imaging and materials science are important examples of such applications. The classical (and fast) methods for computed tomography are based on the assumption that the object of interest remains stationary throughout the scan. However, this assumption is violated when imaging a beating heart or a dynamic process, and this can lead to severe motion artifacts in the computed cross-sectional images.
The purpose of this project is to incorporate the dynamic behaviour of the object of interest into the mathematical reconstruction model that describes the measurement process. Our goal is twofold: 1) we want to demonstrate that the inclusion of an appropriate dynamic model, if available, can overcome some of the shortcomings of existing models and methods, and 2) using techniques from image registration, we seek to estimate the unknown dynamic model that can be included as a prior in the dynamic reconstruction model. These goals are important steps toward improving current dynamic imaging models and methods and may ultimately lead to new applications of X-ray computed tomography.
Supervisors: Martin S. Andersen (Main supervisor), Per Christian Hansen (Co-supervisor)
Section for Scientific Computing
Published as PhD report: Models and methods for dynamic computed tomography