SVoRT: Iterative Transformer for Slice-to-Volume Registration in Fetal Brain MRI

TL;DR

This paper introduces SVoRT, a Transformer-based iterative method for slice-to-volume registration in fetal brain MRI, improving reconstruction accuracy under severe motion artifacts.

Contribution

The paper presents a novel Transformer-based approach that models multiple MRI slices as a sequence and iteratively refines slice-to-volume registration, enhancing fetal brain MRI reconstruction.

Findings

Achieves lower registration error on synthetic data

Improves 3D reconstruction quality under fetal motion

Demonstrates effectiveness on real-world MRI data

Abstract

Volumetric reconstruction of fetal brains from multiple stacks of MR slices, acquired in the presence of almost unpredictable and often severe subject motion, is a challenging task that is highly sensitive to the initialization of slice-to-volume transformations. We propose a novel slice-to-volume registration method using Transformers trained on synthetically transformed data, which model multiple stacks of MR slices as a sequence. With the attention mechanism, our model automatically detects the relevance between slices and predicts the transformation of one slice using information from other slices. We also estimate the underlying 3D volume to assist slice-to-volume registration and update the volume and transformations alternately to improve accuracy. Results on synthetic data show that our method achieves lower registration error and better reconstruction quality compared with existing state-of-the-art methods. Experiments with real-world MRI data are also performed to demonstrate the ability of the proposed model to improve the quality of 3D reconstruction under severe fetal motion.