Image reconstruction in dental tomosynthesis with stationary detector

Abstract

Tomosynthesis is an imaging modality for analyzing anatomical structures. X-ray images from different angles are used for reconstruction in tomosynthesis. In contrast to computed tomography, X-ray tube moves around a restricted pathway for acquisition. Therefore, number of projections are used for reconstruction are less than computed tomography. In this thesis, feasibility of tomosynthesis was explored for a stationary detector system, where x-ray tube had a limited circular movement around isocenter. Algorithms are both used for simulation and physical phantom. X-ray simulation is created based on x-ray attenuation and ray tracing. Physical phantom is produced for acquisition of images with x-ray machine. Observations were made for different conditions in image creation. Effects of the range of circular movement and step size were analyzed. Also, influence of angles between selected plane and axes in image formation were studied. Additionally, effect of selected plane to isocenter distance was experimented. Lastly, artifact and blur removal was applied to achieve better structure identification. Analysis of reconstructed images showed that tomosynthesis could improve the identification of structures. Wider acquisition arcs, artifact and blur removal were the most important factors in getting better images. Planes other than the focused plane caused artifacts and blur in reconstructed image. Therefore, planes with high contrast objects were subtracted from reconstructed plane to improve the visibility of the objects. Histogram analysis also showed the suppression of higher contrast pixels in the reconstructed image.|Keywords : Medical imaging, X-ray imaging, Tomosynthesis, Limited angle imaging, Stationary detector.