Coronary heart disease is the most common cause of death in the developed world. Its main cause is obstruction of the coronary arteries. Today, the evaluation of these stenoses in the interventional lab is performed using C-arm fluoroscopy.

A full 3-D visualisation can improve the clinical assessment. Unfortunately, 3-D reconstruction of the coronary tree from C-arm data is a mathematically ill-posed problem. Due to the long acquisition time of ca. 5 seconds, heart motion blurs the resulting image. There is previous work on the estimation of and compensation for coronary vessel movement, achieving a good visualisation of the coronary morphology. However, heart motion estimation is only approximate, resulting in e.g. uncertainties in the quantitative determination of vessel diameters.

The research goal in this project is the optimisation of the 3-D reconstruction of coronary vasculature towards a quantitative representation. The following aspects are considered:

• Analysis and optimisation of the motion model

• Analysis and improvement of ECG-gated reference image

generation

• Analysis and development of different objective functions

• Analysis and development of optimisation methods for

motion estimation