RemInD: Remembering Anatomical Variations for Interpretable Domain Adaptive Medical Image Segmentation

TL;DR

RemInD introduces a Bayesian framework that memorizes anatomical variations using a domain-agnostic manifold, enabling interpretable and efficient unsupervised domain adaptation for medical image segmentation.

Contribution

It proposes a novel memory-based approach that explicitly captures anatomical variations, improving interpretability and performance in domain adaptation tasks.

Findings

Achieves state-of-the-art results on cardiac and abdominal datasets.

Outperforms existing methods with a single alignment strategy.

Provides explainable segmentation predictions based on anchors and spatial deformation.

Abstract

This work presents a novel Bayesian framework for unsupervised domain adaptation (UDA) in medical image segmentation. While prior works have explored this clinically significant task using various strategies of domain alignment, they often lack an explicit and explainable mechanism to ensure that target image features capture meaningful structural information. Besides, these methods are prone to the curse of dimensionality, inevitably leading to challenges in interpretability and computational efficiency. To address these limitations, we propose RemInD, a framework inspired by human adaptation. RemInD learns a domain-agnostic latent manifold, characterized by several anchors, to memorize anatomical variations. By mapping images onto this manifold as weighted anchor averages, our approach ensures realistic and reliable predictions. This design mirrors how humans develop representative components to understand images and then retrieve component combinations from memory to guide segmentation. Notably, model prediction is determined by two explainable factors: a low-dimensional anchor weight vector, and a spatial deformation. This design facilitates computationally efficient and geometry-adherent adaptation by aligning weight vectors between domains on a probability simplex. Experiments on two public datasets, encompassing cardiac and abdominal imaging, demonstrate the superiority of RemInD, which achieves state-of-the-art performance using a single alignment approach, outperforming existing methods that often rely on multiple complex alignment strategies.