用草图标注提升医学图像分割精度，通过动态教师选择与层级一致性优化
Scribble-Supervised Medical Image Segmentation with Dynamic Teacher Switching and Hierarchical Consistency
- 双教师单学生框架，动态切换更可靠的教师指导
- 在ACDC和MSCMRseg数据集上达到当前最优分割效果
- 适合需要弱监督医学图像分割的临床研究者
草图标注方法可减轻医学图像分割中高昂的标注负担。然而，此类标注固有的稀疏性带来了显著歧义，导致伪标签传播噪声大，阻碍解剖边界的学习。为此，我们提出SDT-Net，一种新型双教师、单学生框架，旨在最大化从这些弱信号中获取的监督质量。该方法包含动态教师切换（DTS）模块，可自适应选择最可靠的教师；该教师通过两个协同机制指导学生：由拾取可靠像素（PRP）机制优化的高置信度伪标签，以及由层级一致性（HiCo）模块强制实施的多级特征对齐。在ACDC和MSCMRseg数据集上的大量实验表明，SDT-Net实现了最先进的性能，生成了更准确且解剖学上合理的分割结果。
原文摘要 · Abstract (English)
Scribble-supervised methods have emerged to mitigate the prohibitive annotation burden in medical image segmentation. However, the inherent sparsity of these annotations introduces significant ambiguity, which results in noisy pseudo-label propagation and hinders the learning of robust anatomical boundaries. To address this challenge, we propose SDT-Net, a novel dual-teacher, single-student framework designed to maximize supervision quality from these weak signals. Our method features a Dynamic Teacher Switching (DTS) module to adaptively select the most reliable teacher. This selected teacher then guides the student via two synergistic mechanisms: high-confidence pseudo-labels, refined by a Pick Reliable Pixels (PRP) mechanism, and multi-level feature alignment, enforced by a Hierarchical Consistency (HiCo) module. Extensive experiments on the ACDC and MSCMRseg datasets demonstrate that SDT-Net achieves state-of-the-art performance, producing more accurate and anatomically plausible segmentation.
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。