用合成数据提升医学图像分割，解决真实与合成图像的语义差异问题。
SRA-Seg: Synthetic to Real Alignment for Semi-Supervised Medical Image Segmentation
- 通过对比学习对齐合成与真实图像的特征分布
- 仅用10%真实标注数据即达89.34%骰子系数
- 适合缺乏标注数据的医疗影像研究者使用
合成数据虽视觉逼真，但在医学图像分割中难以提升性能，因其与真实图像存在语义特征空间差异，导致领域差距。本文提出SRA-Seg框架，通过冻结DINOv2嵌入计算相似性对齐损失，将合成图像特征拉近其最近的真实对应特征。采用软边缘融合生成平滑解剖过渡和连续标签，避免传统拼接增强的硬边界。利用EMA教师模型生成合成图像伪标签，并应用尊重混合区域不确定性的软分割损失。实验表明：仅使用10%标注真实数据与90%合成未标注数据，SRA-Seg在ACDC上取得89.34%的骰子系数，在FIVES上达到84.42%，显著优于现有半监督方法，媲美使用真实未标注数据的方法。
原文摘要 · Abstract (English)
Synthetic data, an appealing alternative to extensive expert-annotated data for medical image segmentation, consistently fails to improve segmentation performance despite its visual realism. The reason being that synthetic and real medical images exist in different semantic feature spaces, creating a domain gap that current semi-supervised learning methods cannot bridge. We propose SRA-Seg, a framework explicitly designed to align synthetic and real feature distributions for medical image segmentation. SRA-Seg introduces a similarity-alignment (SA) loss using frozen DINOv2 embeddings to pull synthetic representations toward their nearest real counterparts in semantic space. We employ soft edge blending to create smooth anatomical transitions and continuous labels, eliminating the hard boundaries from traditional copy-paste augmentation. The framework generates pseudo-labels for synthetic images via an EMA teacher model and applies soft-segmentation losses that respect uncertainty in mixed regions. Our experiments demonstrate strong results: using only 10% labeled real data and 90% synthetic unlabeled data, SRA-Seg achieves 89.34% Dice on ACDC and 84.42% on FIVES, significantly outperforming existing semi-supervised methods and matching the performance of methods using real unlabeled data.
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。