Introduction - If you have any usage issues, please Google them yourself
The registration of pre-operative volumetric datasets to
intra-operative two-dimensional images provides an im-
provedway of verifying patient positionandmedical instru-
ment location. In applications from orthopedics to neuro-
surgery, it has a great value in maintaining up-to-date in-
formation about changes due to intervention. We propose a
mutual information-based registration algorithm which es-
tablishes the proper alignment via a stochastic gradient as-
centstrategy. Ourmaincontributionliesinestimatingprob-
ability density measures of image intensities with a sparse
histogrammingmethodwhichcouldlead to potentialspeed-
up over existing registration procedures and deriving the
gradient estimates required by the maximization procedure.
Experimental results are presented on fluoroscopy and CT
datasets of a real skull, and on a CT-derived dataset of a
real skull, a plastic skull and a plastic lumbar spine seg-
ment.