Learning the Update Operator for 2D/3D Image Registration

TL;DR

This paper introduces a novel deep learning approach that incorporates traditional registration operators to improve the accuracy of 2D/3D image registration in medical imaging.

Contribution

It proposes learning the update step of an iterative registration framework by embedding a known operator, enhancing accuracy over previous deep learning methods.

Findings

Achieved 1.8x improvement in registration accuracy

Successfully integrated traditional operators into deep neural networks

Enhanced robustness and efficiency of 2D/3D registration

Abstract

Image guidance in minimally invasive interventions is usually provided using live 2D X-ray imaging. To enhance the information available during the intervention, the preoperative volume can be overlaid over the 2D images using 2D/3D image registration. Recently, deep learning-based 2D/3D registration methods have shown promising results by improving computational efficiency and robustness. However, there is still a gap in terms of registration accuracy compared to traditional optimization-based methods. We aim to address this gap by incorporating traditional methods in deep neural networks using known operator learning. As an initial step in this direction, we propose to learn the update step of an iterative 2D/3D registration framework based on the Point-to-Plane Correspondence model. We embed the Point-to-Plane Correspondence model as a known operator in our deep neural network and learn the update step for the iterative registration. We show an improvement of 1.8 times in terms of registration accuracy for the update step prediction compared to learning without the known operator.