Interpretable Unsupervised Deformable Image Registration via Confidence-bound Multi-Hop Visual Reasoning

TL;DR

This paper introduces a multi-hop visual reasoning framework for unsupervised deformable image registration that enhances interpretability, robustness, and clinical trust by providing intermediate predictions and confidence estimates.

Contribution

It proposes a novel multi-hop visual chain of reasoning with local refinement and attention mechanisms, enabling transparent and reliable registration of complex anatomical images.

Findings

Achieves competitive accuracy on lung and brain datasets.

Provides interpretable intermediate registration results.

Estimates uncertainty through deformation stability.

Abstract

Unsupervised deformable image registration requires aligning complex anatomical structures without reference labels, making interpretability and reliability critical. Existing deep learning methods achieve considerable accuracy but often lack transparency, leading to error drift and reduced clinical trust. We propose a novel Multi-Hop Visual Chain of Reasoning (VCoR) framework that reformulates registration as a progressive reasoning process. Inspired by the iterative nature of clinical decision-making, each visual reasoning hop integrates a Localized Spatial Refinement (LSR) module to enrich feature representations and a Cross-Reference Attention (CRA) mechanism that leads the iterative refinement process, preserving anatomical consistency. This multi-hop strategy enables robust handling of large deformations and produces a transparent sequence of intermediate predictions with a theoretical bound. Beyond accuracy, our framework offers built-in interpretability by estimating uncertainty via the stability and convergence of deformation fields across hops. Extensive evaluations on two challenging public datasets, DIR-Lab 4D CT (lung) and IXI T1-weighted MRI (brain), demonstrate that VCoR achieves competitive registration accuracy while offering rich intermediate visualizations and confidence measures. By embedding an implicit visual reasoning paradigm, we present an interpretable, reliable, and clinically viable unsupervised medical image registration.