We describe a technique for 4D-reconstruction of coronary arteries from a sequence of monoplane X-ray angiograms. An initial 3D model of coronary centerlines is reconstructed from two appropriate views. A 3D-2D registration framework is formulated in which the model deforms in space to best fit the given angiograms. The 3D motion model is hierarchical and includes rigid, affine and B-spline transformations. The registration is guided by a sum of energy terms which measures the goodness of the 3D-2D mapping and constrains the deformation of the model. The method is tested on three sequences of patient data, each containing 248 frames. The registration time for one frame varies between one and four minutes. © Springer-Verlag Berlin Heidelberg 2005.
CITATION STYLE
Bouattour, S., Arndt, R., & Paulus, D. (2005). 4D reconstruction of coronary arteries from monoplane angiograms. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 3691 LNCS, pp. 724–731). https://doi.org/10.1007/11556121_89
Mendeley helps you to discover research relevant for your work.