In this paper, we propose a hybrid similarity measure for 2D-3D image registration that is a weighted combination of an intensitybased image similarity measure and a point-based measure incorporating a single fiducial marker. We evaluate its accuracy and robustness using gold-standard clinical spine image data. The use of one fiducial marker substantially improves registration accuracy and robustness.
Daniel B. Russakoff, Torsten Rohlfing, Ramin Shahi