Automated segmentation and visualization of the pulmonary vascular tree inspiral CT angiography: An anatomy-oriented approach based on three-dimensional image analysis

A new method for automated segmentation of the pulmonary vascular tree in s piral CT angiography was developed based on 3D image analysis techniques an d anatomic knowledge. For efficient and effective segmentation, an anatomy- oriented approach was introduced, in which several anatomic structures are segmented sequentially and the properties of each segmented structure are u sed for the next step of segmentation and for validation of intermediate re sults. By use of clinical data of 12 patients, parameters for segmentation were analyzed and optimized. The effectiveness of the segmentation method w as evaluated through the visual assessment by comparison between images of the segmentation results by volume rendering and images of maximum intensit y projection of the original volume data.