A general framework for tree segmentation and reconstruction from medical volume data

Abstract

Many anatomical entities like different parts of the arterial system and the airway system are of tree-like structure. State of the art medical imaging systems can acquire 3D volume data of the human body at a resolution that is sufficient for the visualization of these tree structures. We present a general framework for the simultaneous segmentation and reconstruction of the abovementioned entities and apply it to the extraction of coronary arteries from multi detector-row CT data. The coronary artery extraction is evaluated on 9 data-sets with known ground truth for the centerlines of the coronary arteries. © Springer-Verlag Berlin Heidelberg 2004.