---

Intensity-based 2D-3D medical image registration is a special case of the pose estimation problem from computer vision with many applications in medicine. This work presents an overview of the 2D-3D intensity-based image registration problem in the medical domain as well as results from several methods developed to aid in its practice. In particular: 1) Light field rendering techniques from the graphics community are extended to rapidly generate digitally reconstructed radiographs (DRRs). 2) A full 2D-3D registration algorithm using light field DRRs is presented and validated against a real, clinical gold standard. 3) A new, hybrid similarity measure is presented that is a weighted combination of an intensity-based image similarity measure and a point-based measure incorporating a single fiducial marker. 4) Finally, a novel similarity measure called regional mutual information (RMI) is introduced. RMI is an extension of mutual information which incorporates spatial information in a principled way. The additional spatial information helps make its use as a similarity measure much more robust to initial misregistration than standard mutual information.

---

Daniel Russakoff received an A.B. in geophysics from Harvard University and his Ph.D. in computer science from Stanford. His research is in computer vision and pattern recognition in general, and medical image analysis in particular. He has authored numerous papers on topics ranging from stereo vision to medical image registration.

---