2D-3D registration: A step towards image-guided ankle fusion

Abstract

In this paper, we introduce a new framework for registering pre-operative 3D volumetric data to intra-operative 2D images. We are particularly interested in examining the problem of aligning CT volumes to corresponding X-ray images. Our objective is to apply the 2D-3D registration in the field of orthopedics, specifically on ankle fusion surgery. Our framework adopts the shear-warp factorization (SWF) method to generate synthetic 2D images from the given 3D volume. Also, the alignment score is determined based on two novel similarity measures; the exponential correlation (EC) and the pixel-based individual entropy correlation coefficient (IECC). Our framework has been tested on 22 clinical CT datasets.We used different methods to evaluate registration quality of our system. Evaluation results confirm the degree of accuracy and robustness of our proposed framework.