Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Cross-Domain Vessel Segmentation via Latent Similarity Mining and Iterative Co-Optimization

Zhanqiang Guo, Jianjiang Feng, Jie Zhou

Abstract

Retinal vessel segmentation serves as a critical prerequisite for automated diagnosis of retinal pathologies. While recent advances in Convolutional Neural Networks (CNNs) have demonstrated promising performance in this task, significant performance degradation occurs when domain shifts exist between training and testing data. To address these limitations, we propose a novel domain transfer framework that leverages latent vascular similarity across domains and iterative co-optimization of generation and segmentation networks. Specifically, we first pre-train generation networks for source and target domains. Subsequently, the pretrained source-domain conditional diffusion model performs deterministic inversion to establish intermediate latent representations of vascular images, creating domain-agnostic prototypes for target synthesis. Finally, we develop an iterative refinement strategy where segmentation network and generative model undergo mutual optimization through cyclic parameter updating. This co-evolution process enables simultaneous enhancement of cross-domain image synthesis quality and segmentation accuracy. Experiments demonstrate that our framework achieves state-of-the-art performance in cross-domain retinal vessel segmentation, particularly in challenging clinical scenarios with significant modality discrepancies.

Cross-Domain Vessel Segmentation via Latent Similarity Mining and Iterative Co-Optimization

Abstract

Retinal vessel segmentation serves as a critical prerequisite for automated diagnosis of retinal pathologies. While recent advances in Convolutional Neural Networks (CNNs) have demonstrated promising performance in this task, significant performance degradation occurs when domain shifts exist between training and testing data. To address these limitations, we propose a novel domain transfer framework that leverages latent vascular similarity across domains and iterative co-optimization of generation and segmentation networks. Specifically, we first pre-train generation networks for source and target domains. Subsequently, the pretrained source-domain conditional diffusion model performs deterministic inversion to establish intermediate latent representations of vascular images, creating domain-agnostic prototypes for target synthesis. Finally, we develop an iterative refinement strategy where segmentation network and generative model undergo mutual optimization through cyclic parameter updating. This co-evolution process enables simultaneous enhancement of cross-domain image synthesis quality and segmentation accuracy. Experiments demonstrate that our framework achieves state-of-the-art performance in cross-domain retinal vessel segmentation, particularly in challenging clinical scenarios with significant modality discrepancies.

Paper Structure

This paper contains 12 sections, 6 equations, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Method
  3. DDPM and Pre-training
  4. Latent Similarity Mining via DDIM
  5. Iterative Co-Optimization
  6. Experiments and Results
  7. Experimental Setup
  8. Dataset and Evaluation Metrics.
  9. Implementation Details.
  10. Results and Comparison
  11. Iteration Process and Latent Similarity Mining
  12. Conclusion

Figures (4)

  • Figure 1: Cross-modality discrepancy visualization. (a): Samples from FP and OCTA modalities. (b): Intensity distribution histograms. (c): Segmentation performance of Unet ronneberger2015u under in-domain and cross-domain scenarios.
  • Figure 2: Overview of the proposed framework.
  • Figure 3: Illustration of results. The first two rows are the results from the OCTA-500, while the third row is from the ROSE.
  • Figure 4: Performance evolution during iterative optimization.