提出VAOT框架，提升眼底图像增强中血管结构的保真度。
VAOT: Vessel-Aware Optimal Transport for Retinal Fundus Enhancement
- 基于最优传输与血管骨架约束，实现无配对增强
- 在合成退化数据上显著提升血管连通性与末端完整性
- 适合眼科医学图像处理，尤其关注血管结构的场景
彩色眼底摄影（CFP）是诊断和监测视网膜疾病的核心手段，但光照变化等采集差异常导致图像质量下降，亟需鲁棒增强方法。现有无配对增强多基于GAN，但易扭曲关键血管结构，破坏其拓扑与末端完整性。为此，我们提出血管感知最优传输（VAOT）框架，融合最优传输目标与两项结构保持正则项：(i) 基于骨架的损失以维持全局血管连通性；(ii) 终端感知损失以稳定局部血管末端。该机制在无配对设定下引导学习，有效降噪并保留血管结构。在合成退化基准及下游血管与病灶分割任务中，相比多个先进基线方法表现更优。代码已开源：https://github.com/Retinal-Research/VAOT。
原文摘要 · Abstract (English)
Color fundus photography (CFP) is central to diagnosing and monitoring retinal disease, yet its acquisition variability (e.g., illumination changes) often degrades image quality, which motivates robust enhancement methods. Unpaired enhancement pipelines are typically GAN-based, however, they can distort clinically critical vasculature, altering vessel topology and endpoint integrity. Motivated by these structural alterations, we propose Vessel-Aware Optimal Transport (\textbf{VAOT}), a framework that combines an optimal-transport objective with two structure-preserving regularizers: (i) a skeleton-based loss to maintain global vascular connectivity and (ii) an endpoint-aware loss to stabilize local termini. These constraints guide learning in the unpaired setting, reducing noise while preserving vessel structure. Experimental results on synthetic degradation benchmark and downstream evaluations in vessel and lesion segmentation demonstrate the superiority of the proposed methods against several state-of-the art baselines. The code is available at https://github.com/Retinal-Research/VAOT
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。