用大模型辅助生成伪标签，让少标注的医学影像分割更准。
Foundation Model-guided Iteratively Prompting and Pseudo-Labeling for Partially Labeled Medical Image Segmentation
- 用可训练分割网络与冻结大模型协作，迭代生成伪标签。
- 在AMOS数据集上逼近全标注效果，提升显著。
- 适合标注稀缺的临床场景，真实数据验证有效。
全自动医学图像分割在全标注数据下已取得显著进展。然而，因机构临床需求差异及人工标注成本高昂，常出现仅部分器官有标注的“部分标注”问题，导致性能下降。为此，本文提出IPnP框架，通过可训练分割网络（专家）与冻结的基础模型（通才）协同，迭代生成并优化未标注器官的伪标签，逐步恢复完整器官的监督信号。在公开数据集AMOS的模拟部分标注设置下，IPnP持续优于已有方法，接近全标注参考性能。进一步在包含210例头颈部癌患者的私有部分标注数据集上验证，证明其在真实临床场景中的有效性。
原文摘要 · Abstract (English)
Automated medical image segmentation has achieved remarkable progress with fully labeled data. However, site-specific clinical priorities and the high cost of manual annotation often yield scans with only a subset of organs labeled, leading to the partially labeled problem that degrades performance. To address this issue, we propose IPnP, an Iteratively Prompting and Pseudo-labeling framework, for partially labeled medical image segmentation. IPnP iteratively generates and refines pseudo-labels for unlabeled organs through collaboration between a trainable segmentation network (specialist) and a frozen foundation model (generalist), progressively recovering full-organ supervision. On the public dataset AMOS with the simulated partial-label setting, IPnP consistently improves segmentation performance over prior methods and approaches the performance of the fully labeled reference. We further evaluate on a private, partially labeled dataset of 210 head-and-neck cancer patients and demonstrate our effectiveness in real-world clinical settings.
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。