Image processing for computed tomography applications

Segmentation of Vascular Structures in Human Organs In modern medicine, computer-aided image processing plays an increasingly important role. In this thesis, contributions to automated segmentation of pulmonary and hepatic vascular structures from computed tomography angiography data have been made that can be utilized in numerous medical applications including advanced visualization techniques, computer-aided diagnosis, and computerassisted operation planning. Pulmonary vascular tree segmentation is the fundamental basis for different applications, such as the detection and visualization of pulmonary emboli. Within the scope of this work, a novel fuzzy segmentation approach was developed, which can be easily adjusted to the application-specificrequirements on segmentation accuracy. The segmentation of hepatic vascular structures is important for instance for surgical applications such as the planning of oncological resections of liver tumors. A segmentation algorithm has been developed consisting of noise reduction, statistical seed-point detection, and graph-based delineation of hepatic veins. Beyond that, methods for interactive ground truth creation have been introduced. Using the proposed methods, a detailed quantitative evaluation is presented based on 22 pulmonary and 53 hepatic scans demonstrating the effectiveness of the proposed algorithms.