MICDIR: Multi-scale Inverse-consistent Deformable Image Registration using UNetMSS with Self-Constructing Graph Latent

TL;DR

This paper introduces MICDIR, a novel multi-scale, inverse-consistent deformable image registration method using UNetMSS with a self-constructing graph, significantly improving registration accuracy for brain MRIs over existing techniques.

Contribution

The paper extends VoxelMorph by integrating multi-scale supervision, a self-constructing graph latent space, and cycle consistency loss for better large deformation tracking and inverse consistency.

Findings

Achieved higher Dice scores than VoxelMorph and ANTs.

Improved registration accuracy for both intra- and intermodal brain MRI registration.

Demonstrated robustness in capturing large deformations.

Abstract

Image registration is the process of bringing different images into a common coordinate system - a technique widely used in various applications of computer vision, such as remote sensing, image retrieval, and, most commonly, medical imaging. Deep learning based techniques have been applied successfully to tackle various complex medical image processing problems, including medical image registration. Over the years, several image registration techniques have been proposed using deep learning. Deformable image registration techniques such as Voxelmorph have been successful in capturing finer changes and providing smoother deformations. However, Voxelmorph, as well as ICNet and FIRE, do not explicitly encode global dependencies (i.e. the overall anatomical view of the supplied image) and, therefore, cannot track large deformations. In order to tackle the aforementioned problems, this paper extends the Voxelmorph approach in three different ways. To improve the performance in case of small as well as large deformations, supervision of the model at different resolutions has been integrated using a multi-scale UNet. To support the network to learn and encode the minute structural co-relations of the given image-pairs, a self-constructing graph network (SCGNet) has been used as the latent of the multi-scale UNet - which can improve the learning process of the model and help the model to generalise better. And finally, to make the deformations inverse-consistent, cycle consistency loss has been employed. On the task of registration of brain MRIs, the proposed method achieved significant improvements over ANTs and VoxelMorph, obtaining a Dice score of 0.8013 \pm 0.0243 for intramodal and 0.6211 \pm 0.0309 for intermodal, while VoxelMorph achieved 0.7747 \pm 0.0260 and 0.6071 \pm 0.0510, respectively