Registration of 2D Histological Images of Bone Implants with 3D SRμCT Volumes
ISVC '08 Proceedings of the 4th International Symposium on Advances in Visual Computing
WBIR'06 Proceedings of the Third international conference on Biomedical Image Registration
Real-Time 2d/3d deformable registration using metric learning
MCV'12 Proceedings of the Second international conference on Medical Computer Vision: recognition techniques and applications in medical imaging
2D/3D image registration using regression learning
Computer Vision and Image Understanding
| Hi-index | 0.00 |
Registration of a preoperative CT (3D) image to one or more X-rayprojection (2D) images, a special case of the pose estimationproblem, has been attempted in a variety of ways with varyingdegrees of success. Recently, there has been a great deal ofinterest in intensity-based methods. One of the drawbacks to suchmethods is the need to create digitally reconstructed radiographs(DRRs) at each step of the optimization process. DRRs are typicallygenerated by ray casting, an operation that requires O(n3) time,where we assume that n is approximately the size (in voxels) of oneside of the DRR as well as one side of the CT volume. We addressthis issue by extending light field rendering techniques from thecomputer graphics community to generate DRRs instead ofconventional rendered images. Using light fields allows most of thecomputation to be performed in a preprocessing step; after thisprecomputation, very accurate DRRs can be generated in O(n2) time.Another important issue for 2D-3D registration algorithms isvalidation. Previously reported 2D-3D registration algorithms werevalidated using synthetic data or phantoms but not clinical data.We present an intensity-based 2D-3D registration system thatgenerates DRRs using light fields; we validate its performanceusing clinical data with a known gold standard transformation.