Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Leveraging CORAL-Correlation Consistency Network for Semi-Supervised Left Atrium MRI Segmentation

Xinze Li, Runlin Huang, Zhenghao Wu, Bohan Yang, Wentao Fan, Chengzhang Zhu, Weifeng Su

TL;DR

This paper introduces an effective solution, CORAL(Correlation-Aligned)-Correlation Consistency Network (CORN), to capture the global structure shape and local details of Left Atrium, and proposes the Dynamic Feature Pool for the CCM, which utilizes a confidence-based filtering strategy to remove incorrectly selected features and regularize both teacher and student models by constraining the similarity matrix to be consistent.

Abstract

Semi-supervised learning (SSL) has been widely used to learn from both a few labeled images and many unlabeled images to overcome the scarcity of labeled samples in medical image segmentation. Most current SSL-based segmentation methods use pixel values directly to identify similar features in labeled and unlabeled data. They usually fail to accurately capture the intricate attachment structures in the left atrium, such as the areas of inconsistent density or exhibit outward curvatures, adding to the complexity of the task. In this paper, we delve into this issue and introduce an effective solution, CORAL(Correlation-Aligned)-Correlation Consistency Network (CORN), to capture the global structure shape and local details of Left Atrium. Diverging from previous methods focused on each local pixel value, the CORAL-Correlation Consistency Module (CCM) in the CORN leverages second-order statistical information to capture global structural features by minimizing the distribution discrepancy between labeled and unlabeled samples in feature space. Yet, direct construction of features from unlabeled data frequently results in ``Sample Selection Bias'', leading to flawed supervision. We thus further propose the Dynamic Feature Pool (DFP) for the CCM, which utilizes a confidence-based filtering strategy to remove incorrectly selected features and regularize both teacher and student models by constraining the similarity matrix to be consistent. Extensive experiments on the Left Atrium dataset have shown that the proposed CORN outperforms previous state-of-the-art semi-supervised learning methods.

Leveraging CORAL-Correlation Consistency Network for Semi-Supervised Left Atrium MRI Segmentation

TL;DR

This paper introduces an effective solution, CORAL(Correlation-Aligned)-Correlation Consistency Network (CORN), to capture the global structure shape and local details of Left Atrium, and proposes the Dynamic Feature Pool for the CCM, which utilizes a confidence-based filtering strategy to remove incorrectly selected features and regularize both teacher and student models by constraining the similarity matrix to be consistent.

Abstract

Semi-supervised learning (SSL) has been widely used to learn from both a few labeled images and many unlabeled images to overcome the scarcity of labeled samples in medical image segmentation. Most current SSL-based segmentation methods use pixel values directly to identify similar features in labeled and unlabeled data. They usually fail to accurately capture the intricate attachment structures in the left atrium, such as the areas of inconsistent density or exhibit outward curvatures, adding to the complexity of the task. In this paper, we delve into this issue and introduce an effective solution, CORAL(Correlation-Aligned)-Correlation Consistency Network (CORN), to capture the global structure shape and local details of Left Atrium. Diverging from previous methods focused on each local pixel value, the CORAL-Correlation Consistency Module (CCM) in the CORN leverages second-order statistical information to capture global structural features by minimizing the distribution discrepancy between labeled and unlabeled samples in feature space. Yet, direct construction of features from unlabeled data frequently results in ``Sample Selection Bias'', leading to flawed supervision. We thus further propose the Dynamic Feature Pool (DFP) for the CCM, which utilizes a confidence-based filtering strategy to remove incorrectly selected features and regularize both teacher and student models by constraining the similarity matrix to be consistent. Extensive experiments on the Left Atrium dataset have shown that the proposed CORN outperforms previous state-of-the-art semi-supervised learning methods.

Paper Structure

This paper contains 12 sections, 6 equations, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Method
  3. CORAL-Correlation Consistency Module
  4. Dynamic Feature Pool
  5. Experiments and Results
  6. Experiments
  7. Dataset
  8. Implementation Details
  9. Quantitative Evaluation and Comparison
  10. CORAL Loss and Domain Alignment
  11. Ablation Study
  12. Conclusion

Figures (3)

  • Figure 1: Comparative Visualization of Segmentation: Previous Methods vs. CORN. Numerical subscripts denote the results from the model trained with the corresponding number of labeled samples.
  • Figure 2: Overview of our segmentation network. The CORN unifies the Dynamic Feature Pool with a consistency filtering strategy to capture global and local features from both labeled and unlabeled images.
  • Figure 3: The image of CORAL loss curves over training epochs.