3D Reconstruction of Coronary Arteries from Dynamic CAG Image Sequences

A method based on 3D active contour model to sequentially reconstruct vascular axes from single-plane X-ray coronary angiographic (CAG) sequences acquired from two approximately orthogonal views is presented in this paper. The deformation energy of the model combines intrinsic property of the curve and image data with a priori knowledge of coronary arterial configuration and dynamics. 2D extraction, 3D reconstruction and tracking of arterial axes are synchronously implemented through active contour model deforming directly in space. Through introducing a motion model into the energy term, the arterial motion field between consecutive time-points can also be quantitatively estimated.