Springer Verlag, Lecture Notes in Computer Science, p. 1320-1322
DOI: 10.1007/3-540-45468-3_213
Full text: Download
mized segment using an alternative orientation. Finally, the hierarchical structure of the coronary tree is automatically constructed. In the second stage of this method, segmentation and labeling information has to be propagated in the di#erent projections through at least one cardiac cycle. Our approach relies on the hierarchical description of the coronary tree, and the modeling of the arteries by B-snakes with an internal energy derived from the multi-scale model-based preprocessing. A two-step optimization - a global transformation followed by a local optimization of the control points - reduces the e#ects of large displacements, crossings and overlapping arteries. Then, in the third stage, 3D modeling of the coronary artery tree is performed from two segmented projections corresponding roughly to the same cardiac phase. Since they are extracted from two di#erent sequences, heart motion may have occured due to patient's motion or breathing (in the latter, about 10 mm for the inf