DGSSA: Domain generalization with structural and stylistic augmentation for retinal vessel segmentation

2025-01-07

AI Summary

To address poor generalization of retinal vessel segmentation across diverse imaging devices and patient populations, this paper proposes a dual-enhancement framework targeting both anatomical structure and imaging style. First, we introduce a novel spatial colonization algorithm to generate anatomy-consistent vascular structural priors, integrated with an enhanced Pix2Pix model to synthesize training images containing realistic acquisition artifacts. Second, we pioneer the synergistic combination of PixMix photometric mixing and uncertainty-guided pixel-level perturbations to improve style generalization. The method is embedded into U-Net–based segmentation networks and evaluated on four major benchmarks—DRIVE, CHASE_DB1, HRF, and STARE—achieving state-of-the-art performance with an average Dice coefficient improvement of 1.8%. Our approach significantly enhances robustness and generalization to unseen imaging domains, demonstrating superior cross-device and cross-cohort adaptability.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Calibration & Uncertainty Quantification
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Security and Privacy: Privacy-enhancing technologies

Abstract

Retinal vascular morphology is crucial for diagnosing diseases such as diabetes, glaucoma, and hypertension, making accurate segmentation of retinal vessels essential for early intervention. Traditional segmentation methods assume that training and testing data share similar distributions, which can lead to poor performance on unseen domains due to domain shifts caused by variations in imaging devices and patient demographics. This paper presents a novel approach, DGSSA, for retinal vessel image segmentation that enhances model generalization by combining structural and style augmentation strategies. We utilize a space colonization algorithm to generate diverse vascular-like structures that closely mimic actual retinal vessels, which are then used to generate pseudo-retinal images with an improved Pix2Pix model, allowing the segmentation model to learn a broader range of structure distributions. Additionally, we utilize PixMix to implement random photometric augmentations and introduce uncertainty perturbations, thereby enriching stylistic diversity and significantly enhancing the model's adaptability to varying imaging conditions. Our framework has been rigorously evaluated on four challenging datasets-DRIVE, CHASEDB, HRF, and STARE-demonstrating state-of-the-art performance that surpasses existing methods. This validates the effectiveness of our proposed approach, highlighting its potential for clinical application in automated retinal vessel analysis.

Problem

• Retinal Vasculature Segmentation
• Model Generalization
• Disease Diagnosis Assistance

Innovation

• DGSSA
• Cross-domain Generalization
• Retinal Vessel Segmentation