QUIZ: An Arbitrary Volumetric Point Matching Method for Medical Image Registration

TL;DR

This paper introduces QUIZ, a novel medical image registration method that uses CNN and Transformer architectures for accurate voxel point matching, outperforming existing methods especially in large deformation and cross-modality scenarios.

Contribution

The paper presents a new voxel point matching approach for medical image registration that improves accuracy in large deformation and cross-modality cases.

Findings

Achieves smaller deviations than state-of-the-art methods on large deformation datasets.

Surpasses current methods in cross-modality registration accuracy.

Demonstrates robustness in poor image quality conditions.

Abstract

Rigid pre-registration involving local-global matching or other large deformation scenarios is crucial. Current popular methods rely on unsupervised learning based on grayscale similarity, but under circumstances where different poses lead to varying tissue structures, or where image quality is poor, these methods tend to exhibit instability and inaccuracies. In this study, we propose a novel method for medical image registration based on arbitrary voxel point of interest matching, called query point quizzer (QUIZ). QUIZ focuses on the correspondence between local-global matching points, specifically employing CNN for feature extraction and utilizing the Transformer architecture for global point matching queries, followed by applying average displacement for local image rigid transformation. We have validated this approach on a large deformation dataset of cervical cancer patients, with results indicating substantially smaller deviations compared to state-of-the-art methods. Remarkably, even for cross-modality subjects, it achieves results surpassing the current state-of-the-art.