Muliscale Vessel Enhancement Filtering
MICCAI '98 Proceedings of the First International Conference on Medical Image Computing and Computer-Assisted Intervention
A new point matching algorithm for non-rigid registration
Computer Vision and Image Understanding - Special issue on nonrigid image registration
Ray Casting With “On-the-Fly” Region Growing: 3-D Navigation Into Cardiac MSCT Volume
IEEE Transactions on Information Technology in Biomedicine
Hi-index | 0.00 |
This paper describes a method for the characterization of coronary artery motion using multislice computed tomography (MSCT) volume sequences. Coronary trees are first extracted by a spatial vessel tracking method in each volume of MSCT sequence. A point-based matching algorithm, with feature landmarks constraint, is then applied to match the 3-D extracted centerlines between two consecutive instants over a complete cardiac cycle. The transformation functions and correspondence matrices are estimated simultaneously, and allow deformable fitting of the vessels over the volume series. Either point-based or branch-based motion features can be derived. Experiments have been conducted in order to evaluate the performance of the method with amatching error analysis.