SearchMorph:Multi-scale Correlation Iterative Network for Deformable Registration

TL;DR

SearchMorph is an unsupervised multi-scale correlation iterative network that improves deformable image registration accuracy, especially for large deformations and details, by leveraging correlation layers and a deformation field iterator.

Contribution

It introduces a correlation pyramid and deformation field iterator to enhance registration of large deformations and fine details in an unsupervised learning framework.

Findings

Achieves highest registration accuracy among tested methods.

Lowest folding point ratio indicating more realistic deformation fields.

Operates efficiently with short elapsed time.

Abstract

Deformable image registration can obtain dynamic information about images, which is of great significance in medical image analysis. The unsupervised deep learning registration method can quickly achieve high registration accuracy without labels. However, these methods generally suffer from uncorrelated features, poor ability to register large deformations and details, and unnatural deformation fields. To address the issues above, we propose an unsupervised multi-scale correlation iterative registration network (SearchMorph). In the proposed network, we introduce a correlation layer to strengthen the relevance between features and construct a correlation pyramid to provide multi-scale relevance information for the network. We also design a deformation field iterator, which improves the ability of the model to register details and large deformations through the search module and GRU while ensuring that the deformation field is realistic. We use single-temporal brain MR images and multi-temporal echocardiographic sequences to evaluate the model's ability to register large deformations and details. The experimental results demonstrate that the method in this paper achieves the highest registration accuracy and the lowest folding point ratio using a short elapsed time to state-of-the-art.