Unified and Semantically Grounded Domain Adaptation for Medical Image Segmentation

2025-08-12

AI Summary

In medical image segmentation, existing unsupervised domain adaptation (UDA) methods are fragmented across source-available and source-unavailable settings and lack explicit, structured anatomical knowledge modeling, limiting generalizability. This work proposes the first semantic-grounded unified UDA framework: it models anatomical commonality via a domain-invariant probabilistic manifold, enabling natural, handcrafted-strategy-free transfer; and introduces a decoupled anatomical prior learning architecture that jointly leverages manifold-space retrieval and spatial transformation modeling to uniformly support pseudo-labeling, model distillation, and other adaptation mechanisms. Evaluated on multi-center cardiac and abdominal datasets, our method achieves state-of-the-art performance under both settings—under source-unavailable conditions, it approaches the performance of source-available baselines—and additionally enables manifold traversal–based morphological editing and interpretable visualization.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Segmentation
• Reasoning under Uncertainty: Graphical Models

Application Category

• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

Most prior unsupervised domain adaptation approaches for medical image segmentation are narrowly tailored to either the source-accessible setting, where adaptation is guided by source-target alignment, or the source-free setting, which typically resorts to implicit supervision mechanisms such as pseudo-labeling and model distillation. This substantial divergence in methodological designs between the two settings reveals an inherent flaw: the lack of an explicit, structured construction of anatomical knowledge that naturally generalizes across domains and settings. To bridge this longstanding divide, we introduce a unified, semantically grounded framework that supports both source-accessible and source-free adaptation. Fundamentally distinct from all prior works, our framework's adaptability emerges naturally as a direct consequence of the model architecture, without the need for any handcrafted adaptation strategies. Specifically, our model learns a domain-agnostic probabilistic manifold as a global space of anatomical regularities, mirroring how humans establish visual understanding. Thus, the structural content in each image can be interpreted as a canonical anatomy retrieved from the manifold and a spatial transformation capturing individual-specific geometry. This disentangled, interpretable formulation enables semantically meaningful prediction with intrinsic adaptability. Extensive experiments on challenging cardiac and abdominal datasets show that our framework achieves state-of-the-art results in both settings, with source-free performance closely approaching its source-accessible counterpart, a level of consistency rarely observed in prior works. Beyond quantitative improvement, we demonstrate strong interpretability of the proposed framework via manifold traversal for smooth shape manipulation.

Problem

• Unifies source-accessible and source-free medical image segmentation adaptation
• Lacks explicit anatomical knowledge construction across domains
• Proposes domain-agnostic probabilistic manifold for interpretable predictions

Innovation

• Unified framework for source-accessible and source-free adaptation
• Domain-agnostic probabilistic manifold for anatomical regularities
• Disentangled interpretable formulation enabling adaptable predictions