用自监督+伪标签，让少标注的CBCT牙体分割更准更快
U-Mamba2-SSL for Semi-Supervised Tooth and Pulp Segmentation in CBCT
- 基于U-Mamba2构建半监督框架，分三阶段训练提升泛化能力
- 在隐藏测试集上达0.917 DSC，STSR 2025挑战赛任务1第一
- 适合医疗影像少样本场景，尤其对标注成本高的牙科应用
锥形束计算机断层扫描（CBCT）中牙体与牙髓的精确分割对治疗规划和诊断至关重要。然而该过程需大量专业经验且耗时极长，亟需能有效利用未标注数据的自动化算法。本文提出U-Mamba2-SSL，一种基于U-Mamba2模型的新型半监督学习框架，采用多阶段训练策略：首先通过破坏性自编码器进行自监督预训练；接着利用一致性正则化，引入输入与特征扰动以保证模型输出稳定；最后采用伪标签策略并降低损失权重，减轻错误标签影响。U-Mamba2-SSL在隐藏测试集上平均得分0.789，DSC达0.917，夺得STSR 2025挑战赛任务1第一名。代码已公开于https://github.com/zhiqin1998/UMamba2。
原文摘要 · Abstract (English)
Accurate segmentation of teeth and pulp in Cone-Beam Computed Tomography (CBCT) is vital for clinical applications like treatment planning and diagnosis. However, this process requires extensive expertise and is exceptionally time-consuming, highlighting the critical need for automated algorithms that can effectively utilize unlabeled data. In this paper, we propose U-Mamba2-SSL, a novel semi-supervised learning framework that builds on the U-Mamba2 model and employs a multi-stage training strategy. The framework first pre-trains U-Mamba2 in a self-supervised manner using a disruptive autoencoder. It then leverages unlabeled data through consistency regularization, where we introduce input and feature perturbations to ensure stable model outputs. Finally, a pseudo-labeling strategy is implemented with a reduced loss weighting to minimize the impact of potential errors. U-Mamba2-SSL achieved an average score of 0.789 and a DSC of 0.917 on the hidden test set, achieving first place in Task 1 of the STSR 2025 challenge. The code is available at https://github.com/zhiqin1998/UMamba2.
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。