An automated matching algorithm for dual orthogonal fluoroscopy

Abstract

The current method of studying in vivo kinematics of human joints is a tedious and time consuming task. Current techniques require the segmentation of hundreds of raster images, often by hand, and registration of three-dimensional models with two-dimensional contours. Automation of these processes would greatly accelerate research in this field. This research presents an automated method for recovering the pose of a three dimensional model from planar contours. Validation of the algorithm and discussion of its performance and applicability are detailed herein.