Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

GIGP: A Global Information Interacting and Geometric Priors Focusing Framework for Semi-supervised Medical Image Segmentation

Lianyuan Yu, Xiuzhen Guo, Ji Shi, Hongxiao Wang, Hongwei Li

TL;DR

The paper tackles semi-supervised medical image segmentation where labeled data are scarce and a distribution shift exists between labeled and unlabeled data. It introduces GIGP, a global-perspective framework built on a Mean Teacher backbone that combines GIIM, GMAM, and GGPC, with the training objective $L = L_s + \gamma_1 L_c + \gamma_2 L_{gmc}$. GIIM enables cross-distribution feature learning via four directional Mamba-based interactions; GMAM leverages second-order geometric moments for multi-view, multi-scale geometric guidance; GGPC imposes periodic sine-wave perturbations on 3D coordinates to simulate organ dynamics. On NIH Pancreas and Left Atrium datasets, GIGP achieves state-of-the-art Dice and Jaccard under limited labels, with ablation analyses confirming the contribution of each component.

Abstract

Semi-supervised learning enhances medical image segmentation by leveraging unlabeled data, reducing reliance on extensive labeled datasets. On the one hand, the distribution discrepancy between limited labeled data and abundant unlabeled data can hinder model generalization. Most existing methods rely on local similarity matching, which may introduce bias. In contrast, Mamba effectively models global context with linear complexity, learning more comprehensive data representations. On the other hand, medical images usually exhibit consistent anatomical structures defined by geometric features. Most existing methods fail to fully utilize global geometric priors, such as volumes, moments etc. In this work, we introduce a global information interaction and geometric priors focus framework (GIGP). Firstly, we present a Global Information Interaction Mamba module to reduce distribution discrepancy between labeled and unlabeled data. Secondly, we propose a Geometric Moment Attention Mechanism to extract richer global geometric features. Finally, we propose Global Geometric Perturbation Consistency to simulate organ dynamics and geometric variations, enhancing the ability of the model to learn generalized features. The superior performance on the NIH Pancreas and Left Atrium datasets demonstrates the effectiveness of our approach.

GIGP: A Global Information Interacting and Geometric Priors Focusing Framework for Semi-supervised Medical Image Segmentation

TL;DR

The paper tackles semi-supervised medical image segmentation where labeled data are scarce and a distribution shift exists between labeled and unlabeled data. It introduces GIGP, a global-perspective framework built on a Mean Teacher backbone that combines GIIM, GMAM, and GGPC, with the training objective . GIIM enables cross-distribution feature learning via four directional Mamba-based interactions; GMAM leverages second-order geometric moments for multi-view, multi-scale geometric guidance; GGPC imposes periodic sine-wave perturbations on 3D coordinates to simulate organ dynamics. On NIH Pancreas and Left Atrium datasets, GIGP achieves state-of-the-art Dice and Jaccard under limited labels, with ablation analyses confirming the contribution of each component.

Abstract

Semi-supervised learning enhances medical image segmentation by leveraging unlabeled data, reducing reliance on extensive labeled datasets. On the one hand, the distribution discrepancy between limited labeled data and abundant unlabeled data can hinder model generalization. Most existing methods rely on local similarity matching, which may introduce bias. In contrast, Mamba effectively models global context with linear complexity, learning more comprehensive data representations. On the other hand, medical images usually exhibit consistent anatomical structures defined by geometric features. Most existing methods fail to fully utilize global geometric priors, such as volumes, moments etc. In this work, we introduce a global information interaction and geometric priors focus framework (GIGP). Firstly, we present a Global Information Interaction Mamba module to reduce distribution discrepancy between labeled and unlabeled data. Secondly, we propose a Geometric Moment Attention Mechanism to extract richer global geometric features. Finally, we propose Global Geometric Perturbation Consistency to simulate organ dynamics and geometric variations, enhancing the ability of the model to learn generalized features. The superior performance on the NIH Pancreas and Left Atrium datasets demonstrates the effectiveness of our approach.

Paper Structure

This paper contains 9 sections, 9 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Overview
  4. Global Information Interaction Mamba
  5. Geometric Moment Attention Mechanism
  6. Global Geometric Perturbation Consistency
  7. Experiments
  8. Conclusion
  9. Acknowledgments

Figures (4)

  • Figure 1: The overview of GIGP. $L_{E_k,gmc}$ and $L_{D_k,gmc}$ represent the the moment losses of the k-th layer of the encoder and decoder, respectively. $GM^{h}$, $GM^{w}$, and $GM^{l}$ represent computing geometric moments in three dimensions, respectively.
  • Figure 2: The overview of Geometric Information Propagation Mamba. GSC is the gated spatial convolution module b30. LayerNorm denotes the layer normalization, and MLP represents the multiple layers perception layer to enrich the feature representation. IIM is the Information Interaction Mamba.
  • Figure 3: Results of qualitative comparison on Pancreas dataset under 20$\%$ labeled data setting. GT represents the ground truth.
  • Figure 4: Results of qualitative comparison on LA dataset under 10$\%$ labeled data setting. GT represents the ground truth.