Contrastive Graph Modeling for Cross-Domain Few-Shot Medical Image Segmentation

TL;DR

This work introduces Contrastive Graph Modeling (C-Graph) for cross-domain few-shot medical image segmentation, leveraging a Structural Prior Graph to capture domain-transferable anatomical structure and a Subgraph Matching Decoding mechanism to exploit node connectivity for prediction. A Confusion-minimizing Node Contrast loss reduces semantic ambiguity in graph space, enabling robust cross-domain performance while preserving source-domain accuracy. Extensive experiments across four medical datasets show state-of-the-art results on cross-domain tasks and strong source-domain performance, with thorough ablations demonstrating the contributions of SPG, SMD, and CNC. The approach offers a principled framework for encoding domain-agnostic structure in medical images and achieving scalable generalization with limited labeled data.

Abstract

Cross-domain few-shot medical image segmentation (CD-FSMIS) offers a promising and data-efficient solution for medical applications where annotations are severely scarce and multimodal analysis is required. However, existing methods typically filter out domain-specific information to improve generalization, which inadvertently limits cross-domain performance and degrades source-domain accuracy. To address this, we present Contrastive Graph Modeling (C-Graph), a framework that leverages the structural consistency of medical images as a reliable domain-transferable prior. We represent image features as graphs, with pixels as nodes and semantic affinities as edges. A Structural Prior Graph (SPG) layer is proposed to capture and transfer target-category node dependencies and enable global structure modeling through explicit node interactions. Building upon SPG layers, we introduce a Subgraph Matching Decoding (SMD) mechanism that exploits semantic relations among nodes to guide prediction. Furthermore, we design a Confusion-minimizing Node Contrast (CNC) loss to mitigate node ambiguity and subgraph heterogeneity by contrastively enhancing node discriminability in the graph space. Our method significantly outperforms prior CD-FSMIS approaches across multiple cross-domain benchmarks, achieving state-of-the-art performance while simultaneously preserving strong segmentation accuracy on the source domain.

Figures (9)

• Figure 1: (a) Our motivation. Despite significant appearance shifts, medical images exhibit high structural consistency across domains. (b) Previous methods naively focus on filtering out domain-specific information to improve generalization, while overlooking feature collapse, limiting cross-domain performance and severely degrading source-domain accuracy. (c) Our method models domain-consistent structure using graphs and employs contrastive learning to reduce node semantic confusion, achieving both superior in- and cross-domain performance.
• Figure 2: Overview of our proposed method. Here, $\mathcal{G}^i$ denotes the output graph of the $i$-th SPG layer.
• Figure 3: Visual and quantitative comparison of segmentation results. Top-left: CT $\to$ MRI; top-right: MRI $\to$ CT; bottom-left: LGE $\to$ b-SSFP; bottom-right: b-SSFP $\to$ LGE.
• Figure 4: Iteration of entropy maps for prediction results. As training progresses, confusion decreases from high (bright yellow) to low (dark red), indicating that $\mathcal{L}_{cnc}$ significantly improves node discriminability.
• Figure 5: Visualization of the graphs constructed from intermediate features, consistent with the graph structure employed in the model pipeline. (a) Target domain images and zoomed-in view of the yellow box regions. (b) Graph representations over feature map coordinates. (c) UMAP UMAP visualization of the graphs in feature space. All graphs are constructed with $k=9$. Nodes with the same color denote the same entity across the graphs in (b) and (c). The edges collectively reflect anatomical semantics in the target-domain image, indicating that the graph effectively captures its underlying structural patterns. Best viewed in color and with zoom.