Efficient Prealignment of CT Scans for Registration through a Bodypart Regressor

TL;DR

This paper introduces a CNN-based bodypart regressor to facilitate fast and robust prealignment of CT scans, improving registration efficiency by leveraging relative height scores for quick alignment across timepoints.

Contribution

It presents a novel application of CNNs as a bodypart regressor specifically for prealigning CT scans, enhancing registration speed and robustness.

Findings

Fast prealignment compared to iterative methods

Robustness across different timepoints

Preliminary results confirm effectiveness

Abstract

Convolutional neural networks have not only been applied for classification of voxels, objects, or images, for instance, but have also been proposed as a bodypart regressor. We pick up this underexplored idea and evaluate its value for registration: A CNN is trained to output the relative height within the human body in axial CT scans, and the resulting scores are used for quick alignment between different timepoints. Preliminary results confirm that this allows both fast and robust prealignment compared with iterative approaches.