Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

FAMNet: Frequency-aware Matching Network for Cross-domain Few-shot Medical Image Segmentation

Yuntian Bo, Yazhou Zhu, Lunbo Li, Haofeng Zhang

TL;DR

FAMNet tackles cross-domain few-shot medical image segmentation by decoupling foreground features into frequency bands and aligning support–query representations in a frequency-aware manner. The Frequency-aware Matching (FAM) module mitigates intra-domain bias and inter-domain shifts by band-wise matching and treating domain-variant versus domain-invariant bands differently, while the Multi-Spectral Fusion (MSF) module cross-estimates information across bands to suppress domain-variant information. Built on a ResNet-50 backbone with a Coarse Prediction Generation (CPG) stage and trained with dual BCE losses, the approach achieves state-of-the-art Dice scores on three cross-domain datasets (Cross-Modality, Cross-Sequence, Cross-Institution) in 1-shot 1-way settings, demonstrating strong generalization under modality shifts. This work advances CD-FSMIS with a practical, frequency-domain strategy that reduces reliance on domain-specific features and improves segmentation robustness in medical imaging.

Abstract

Existing few-shot medical image segmentation (FSMIS) models fail to address a practical issue in medical imaging: the domain shift caused by different imaging techniques, which limits the applicability to current FSMIS tasks. To overcome this limitation, we focus on the cross-domain few-shot medical image segmentation (CD-FSMIS) task, aiming to develop a generalized model capable of adapting to a broader range of medical image segmentation scenarios with limited labeled data from the novel target domain. Inspired by the characteristics of frequency domain similarity across different domains, we propose a Frequency-aware Matching Network (FAMNet), which includes two key components: a Frequency-aware Matching (FAM) module and a Multi-Spectral Fusion (MSF) module. The FAM module tackles two problems during the meta-learning phase: 1) intra-domain variance caused by the inherent support-query bias, due to the different appearances of organs and lesions, and 2) inter-domain variance caused by different medical imaging techniques. Additionally, we design an MSF module to integrate the different frequency features decoupled by the FAM module, and further mitigate the impact of inter-domain variance on the model's segmentation performance. Combining these two modules, our FAMNet surpasses existing FSMIS models and Cross-domain Few-shot Semantic Segmentation models on three cross-domain datasets, achieving state-of-the-art performance in the CD-FSMIS task.

FAMNet: Frequency-aware Matching Network for Cross-domain Few-shot Medical Image Segmentation

TL;DR

FAMNet tackles cross-domain few-shot medical image segmentation by decoupling foreground features into frequency bands and aligning support–query representations in a frequency-aware manner. The Frequency-aware Matching (FAM) module mitigates intra-domain bias and inter-domain shifts by band-wise matching and treating domain-variant versus domain-invariant bands differently, while the Multi-Spectral Fusion (MSF) module cross-estimates information across bands to suppress domain-variant information. Built on a ResNet-50 backbone with a Coarse Prediction Generation (CPG) stage and trained with dual BCE losses, the approach achieves state-of-the-art Dice scores on three cross-domain datasets (Cross-Modality, Cross-Sequence, Cross-Institution) in 1-shot 1-way settings, demonstrating strong generalization under modality shifts. This work advances CD-FSMIS with a practical, frequency-domain strategy that reduces reliance on domain-specific features and improves segmentation robustness in medical imaging.

Abstract

Existing few-shot medical image segmentation (FSMIS) models fail to address a practical issue in medical imaging: the domain shift caused by different imaging techniques, which limits the applicability to current FSMIS tasks. To overcome this limitation, we focus on the cross-domain few-shot medical image segmentation (CD-FSMIS) task, aiming to develop a generalized model capable of adapting to a broader range of medical image segmentation scenarios with limited labeled data from the novel target domain. Inspired by the characteristics of frequency domain similarity across different domains, we propose a Frequency-aware Matching Network (FAMNet), which includes two key components: a Frequency-aware Matching (FAM) module and a Multi-Spectral Fusion (MSF) module. The FAM module tackles two problems during the meta-learning phase: 1) intra-domain variance caused by the inherent support-query bias, due to the different appearances of organs and lesions, and 2) inter-domain variance caused by different medical imaging techniques. Additionally, we design an MSF module to integrate the different frequency features decoupled by the FAM module, and further mitigate the impact of inter-domain variance on the model's segmentation performance. Combining these two modules, our FAMNet surpasses existing FSMIS models and Cross-domain Few-shot Semantic Segmentation models on three cross-domain datasets, achieving state-of-the-art performance in the CD-FSMIS task.

Paper Structure

This paper contains 20 sections, 19 equations, 2 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Few-shot Medical Image Segmentation
  4. Cross-domain Few-shot Semantic Segmentation
  5. Methodology
  6. Problem Setting
  7. Overall Architecture
  8. Feature Extraction
  9. Coarse Prediction Generation (CPG) Module
  10. Frequency-Aware Matching (FAM) Module
  11. Multi-Spectral Fusion (MSF) Module
  12. Objective Function
  13. Experiments
  14. Datasets
  15. Implementation Details
  16. ...and 5 more sections

Figures (2)

  • Figure 1: Motivation of the proposed method. (a) CT and MRI scans in the spatial and frequency domains. Frequency spectra are processed using a Hamming window HAMMING and are center-shifted. (b) Quantitative metrics for the similarity of CT and MRI in the spatial and frequency domains using structural similarity index measure (SSIM) SSIM and normalized mean square error (NMSE). Metrics are calculated using registered images.
  • Figure 2: The overall architecture of our method, consists of three main technical components: the Coarse Prediction Generation (CPG) module, the Frequency-aware Matching (FAM) module, and the Multi-Spectral Fusion (MSF) module. Note that in ABM, JSM denotes joint space matching. In the case of DAFBs, the attention matrix is directly utilized for attention weighting. Conversely, for DSFBs, an element-wise subtraction is applied prior to the weighting process.