Abstract
The original idea underlying this thesis was to show whether and to what extent it is feasible to use a surgical C-arm for volume reconstruction in the field of medical imaging. Surgical C-arms are widely used as intra-operational imaging tools, for instance for navigation or verification. They provide the possibility to easily obtain X-ray images of a patient from an arbitrary viewing angle. These features make them a promising candidate for the use in volume reconstruction, since this requires the recording of a number of images from many directions. In the last years, several approaches with different kinds of C-arms have been undertaken by different research groups or companies. Since the three-dimensional imaging of patients requires an accurate knowledge of the projection geometry, the larger, more rigid, and motorised vascular G-arm systems have been the gantries of choice. These C-arms are usually heavy and ceiling-mounted. They require a large room to be placed and moved in. Surgical C-arms, on the other hand, are less stable, but movable and flexible. The feasibility of using a movable surgical C-arm for volume reconstruction of its linear attenuation coefficients will be evaluated in this thesis. To this end, system calibration techniques, reconstruction methods, visualisation techniques, and possible medical applications have been explored. It will be shown, which extend of calibration is necessary for a surgical C- arm to provide the quality of data needed for volume reconstruction. The reconstructions are performed on the basis of a standard algorithm due to Feldkamp et al. Results are presented as maximum-intensity-projections (MIPs), volume slices as in classic CT, and volume-rendered images of the reconstructed volumes. These images could be used for the diagnosis of bone-related injuries or diseases, the planning and/or verification of the placement of pins supporting the recovery of splintered bones or the modelling of prostheses with a higher precision than today. The work includes the development of new interpolative calibration methods that enable the use of arbitrary system trajectories in future C-arm gantries.