Deformable Medical Image Registration with Effective Anatomical Structure Representation and Divide-and-Conquer Network

2025-06-23

AI Summary

Existing unsupervised deformable medical image registration (DMIR) methods neglect anatomical structure representation, while weakly supervised approaches over-rely on manual labels. To address these limitations, this paper proposes a fully unsupervised “divide-and-conquer” registration framework. Its core contributions are: (1) adaptive extraction of multi-intensity effective regions-of-interest (ROIs) via Gaussian mixture modeling to enhance anatomical representation; and (2) a channel-wise convolutional network (DCN) that enables ROI-level independent alignment and subsequent fusion into a globally consistent deformation field. The method requires no human annotations. Evaluated on multiple MRI and CT datasets, it significantly outperforms VoxelMorph, achieving Dice score improvements of 5.75–13.01%. It delivers superior registration accuracy and enhanced control over deformation smoothness and plausibility.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Unsupervised & Self-Supervised Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Effective representation of Regions of Interest (ROI) and independent alignment of these ROIs can significantly enhance the performance of deformable medical image registration (DMIR). However, current learning-based DMIR methods have limitations. Unsupervised techniques disregard ROI representation and proceed directly with aligning pairs of images, while weakly-supervised methods heavily depend on label constraints to facilitate registration. To address these issues, we introduce a novel ROI-based registration approach named EASR-DCN. Our method represents medical images through effective ROIs and achieves independent alignment of these ROIs without requiring labels. Specifically, we first used a Gaussian mixture model for intensity analysis to represent images using multiple effective ROIs with distinct intensities. Furthermore, we propose a novel Divide-and-Conquer Network (DCN) to process these ROIs through separate channels to learn feature alignments for each ROI. The resultant correspondences are seamlessly integrated to generate a comprehensive displacement vector field. Extensive experiments were performed on three MRI and one CT datasets to showcase the superior accuracy and deformation reduction efficacy of our EASR-DCN. Compared to VoxelMorph, our EASR-DCN achieved improvements of 10.31% in the Dice score for brain MRI, 13.01% for cardiac MRI, and 5.75% for hippocampus MRI, highlighting its promising potential for clinical applications. The code for this work will be released upon acceptance of the paper.

Problem

• Improves deformable medical image registration via ROI representation
• Addresses limitations of unsupervised and weakly-supervised DMIR methods
• Achieves label-free independent alignment of anatomical ROIs

Innovation

• Gaussian mixture model for ROI representation
• Divide-and-Conquer Network for independent alignment
• Label-free multi-ROI feature integration