文献速递牙体牙髓病

自监督学习提升有限标注根尖片分割

Self-supervised learning enhances periapical films segmentation with limited labeled data.
PubMed 文献 · J Dent · 2025-10-09

本研究提出两阶段自监督学习框架:先在74,292张无标注根尖片上自监督预训练ViT,再用229张标注片微调Mask2Former头分割七种牙科结构。平均Dice系数74.77%,显著优于U-Net和DeepLabV3+(33.53%-41.55%),也优于MAE、MoCov3和BEiTv3。

要点

  • 两阶段自监督框架仅需229张标注片即可分割七种牙科结构。
  • 平均Dice系数74.77%,显著优于传统全监督模型和前沿自监督方法。
  • 该方法为资源有限场景下开发AI辅助诊断工具提供可行路径。

核心结论

  • 两阶段领域特定自监督框架能有效学习稳健解剖特征,在极少标注下实现准确可靠的根尖片分割。
  • 该方法优于传统全监督模型和前沿自监督方法,解决了医学影像标注数据稀缺的挑战。

研究设计与样本

  • 研究提出两阶段框架:自监督预训练和微调。
  • 自监督预训练:ViT以DINOv2(在142M自然图像LVD-142M上预训练)权重初始化,在74,292张无标注根尖片上用学生-教师对比学习进一步预训练。
  • 微调:领域适配ViT与Mask2Former头在仅229张标注片上微调,分割七种关键牙科结构(牙、牙髓、牙冠、充填物、根管充填物、龋齿、根尖周病变)。

主要发现

  • 领域适配自监督方法平均Dice系数74.77%,显著优于U-Net和DeepLabV3+的33.53%-41.55%(相对提升80%-123%)。
  • 交叉验证中DINOv2方法(74.77±1.87%)优于MAE(72.53±1.90%)、MoCov3(65.92±1.68%)和BEiTv3(65.17±1.77%)。
  • 该方法以统计学显著性优于其监督Mask2Former对应方法(p<0.01)。

临床怎么用

  • 该方法为开发AI辅助诊断工具提供可行路径,可在资源受限牙科实践中通过一致分割提高诊断准确性并减少手动分析时间。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

OBJECTIVES: To overcome reliance on large-scale, costly labeled datasets and annotation variability for accurate periapical film segmentation. This study develops a self-supervised learning framework requiring limited labeled data, enhancing practical applicability while reducing extensive manual annotation efforts. METHODS: This research proposes a two-stage framework: 1) Self-supervised pre-training. A Vision Transformer (ViT), initialized with weights from the DINOv2 model pre-trained on 142 M natural images (LVD-142 M), undergoes further self-supervised pre-training on our dataset of 74,292 unlabeled periapical films using student-teacher contrastive learning. 2) Fine-tuning adapts these features for segmentation. The domain-adapted ViT is fine-tuned with a Mask2Former head on only 229 labeled films for segmenting seven critical dental structures (tooth, pulp, crown, fillings, root canal fillings, caries, periapical lesions). RESULTS: The domain-adapted self-supervised method significantly outperformed traditional fully-supervised models like U-Net and DeepLabV3+ (average Dice coefficient: 74.77 % vs 33.53 %-41.55 %; 80 %-123 % relative improvement). Comprehensive comparison with cutting-edge SSL methods through cross-validation demonstrated the superiority of our DINOv2-based approach (74.77 ± 1.87 %) over MAE (72.53 ± 1.90 %), MoCov3 (65.92 ± 1.68 %) and BEiTv3 (65.17 ± 1.77 %). The method surpassed its supervised Mask2Former counterparts with statistical significance (p < 0.01). CONCLUSIONS: This two-stage, domain-specific self-supervised framework effectively learns robust anatomical features. It enables accurate, reliable periapical film segmentation using very limited annotations. The approach addresses the challenge of labeled data scarcity in medical imaging. CLINICAL SIGNIFICANCE: This approach provides a feasible pathway for developing AI-assisted diagnostic tools. It can improve diagnostic accuracy through consistent segmentation and enhance workflow efficiency by reducing manual analysis time, especially in resource-constrained dental practices.

来源:J Dent,查看原文。版权归原作者所有。

引用

  1. Hu M, Zhang Q, Wei Z, et al. Self-supervised learning enhances periapical films segmentation with limited labeled data. J Dent. 2025. PMID: 41067573. doi:10.1016/j.jdent.2025.106150

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。