Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Domain-invariant Mixed-domain Semi-supervised Medical Image Segmentation with Clustered Maximum Mean Discrepancy Alignment

Ba-Thinh Lam, Thanh-Huy Nguyen, Hoang-Thien Nguyen, Quang-Khai Bui-Tran, Nguyen Lan Vi Vu, Phat K. Huynh, Ulas Bagci, Min Xu

TL;DR

This work tackles mixed-domain semi-supervised medical image segmentation where domain labels are unknown and data come from multiple scanners. It introduces a domain-invariant framework that combines a Copy-Paste Mechanism (CPM) to expand cross-domain training with a Cluster Maximum Mean Discrepancy (CMMD) block to cluster unlabeled features and align them with labeled anchors, all within a teacher–student architecture. CMMD comprises a Clustering Module (via HDBSCAN) to identify pseudo-domain centroids and an MMD-based Domain-Aware Module to minimize divergence between clusters and the labeled anchor, applied to encoder layers to remove domain signals before decoding. Experiments on Fundus and M&Ms show the approach consistently outperforms SSL and DA baselines, demonstrating strong robustness under limited labels and multiple unknown domain gaps, with practical implications for real-world mixed-domain deployments.

Abstract

Deep learning has shown remarkable progress in medical image semantic segmentation, yet its success heavily depends on large-scale expert annotations and consistent data distributions. In practice, annotations are scarce, and images are collected from multiple scanners or centers, leading to mixed-domain settings with unknown domain labels and severe domain gaps. Existing semi-supervised or domain adaptation approaches typically assume either a single domain shift or access to explicit domain indices, which rarely hold in real-world deployment. In this paper, we propose a domain-invariant mixed-domain semi-supervised segmentation framework that jointly enhances data diversity and mitigates domain bias. A Copy-Paste Mechanism (CPM) augments the training set by transferring informative regions across domains, while a Cluster Maximum Mean Discrepancy (CMMD) block clusters unlabeled features and aligns them with labeled anchors via an MMD objective, encouraging domain-invariant representations. Integrated within a teacher-student framework, our method achieves robust and precise segmentation even with very few labeled examples and multiple unknown domain discrepancies. Experiments on Fundus and M&Ms benchmarks demonstrate that our approach consistently surpasses semi-supervised and domain adaptation methods, establishing a potential solution for mixed-domain semi-supervised medical image segmentation.

Domain-invariant Mixed-domain Semi-supervised Medical Image Segmentation with Clustered Maximum Mean Discrepancy Alignment

TL;DR

This work tackles mixed-domain semi-supervised medical image segmentation where domain labels are unknown and data come from multiple scanners. It introduces a domain-invariant framework that combines a Copy-Paste Mechanism (CPM) to expand cross-domain training with a Cluster Maximum Mean Discrepancy (CMMD) block to cluster unlabeled features and align them with labeled anchors, all within a teacher–student architecture. CMMD comprises a Clustering Module (via HDBSCAN) to identify pseudo-domain centroids and an MMD-based Domain-Aware Module to minimize divergence between clusters and the labeled anchor, applied to encoder layers to remove domain signals before decoding. Experiments on Fundus and M&Ms show the approach consistently outperforms SSL and DA baselines, demonstrating strong robustness under limited labels and multiple unknown domain gaps, with practical implications for real-world mixed-domain deployments.

Abstract

Deep learning has shown remarkable progress in medical image semantic segmentation, yet its success heavily depends on large-scale expert annotations and consistent data distributions. In practice, annotations are scarce, and images are collected from multiple scanners or centers, leading to mixed-domain settings with unknown domain labels and severe domain gaps. Existing semi-supervised or domain adaptation approaches typically assume either a single domain shift or access to explicit domain indices, which rarely hold in real-world deployment. In this paper, we propose a domain-invariant mixed-domain semi-supervised segmentation framework that jointly enhances data diversity and mitigates domain bias. A Copy-Paste Mechanism (CPM) augments the training set by transferring informative regions across domains, while a Cluster Maximum Mean Discrepancy (CMMD) block clusters unlabeled features and aligns them with labeled anchors via an MMD objective, encouraging domain-invariant representations. Integrated within a teacher-student framework, our method achieves robust and precise segmentation even with very few labeled examples and multiple unknown domain discrepancies. Experiments on Fundus and M&Ms benchmarks demonstrate that our approach consistently surpasses semi-supervised and domain adaptation methods, establishing a potential solution for mixed-domain semi-supervised medical image segmentation.
Paper Structure (14 sections, 12 equations, 3 figures, 4 tables)

This paper contains 14 sections, 12 equations, 3 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Problem Definition
  4. Copy-Paste Mechanism
  5. Cluster Maximum Mean Discrepancy Block
  6. Clustering Module
  7. MMD-based Domain-Aware Module
  8. Domain Loss
  9. Final Loss
  10. Experimental Results
  11. Datasets & Metrics
  12. Results
  13. Ablation & Analysis
  14. Conclusion

Figures (3)

  • Figure 1: Feature visualizations of the student encoder’s last layer on the Fundus dataset using UMAP. Domains are shown in different colors, and the model was trained with labeled samples from Domain 4. Silhouette Score assesses cluster quality, with higher scores indicating more separated clusters.
  • Figure 2: Overall illustration of Our Framework.
  • Figure 3: Illustration of CMMD Block.