RAIL：面向CBCT半监督牙齿分割的区域感知指导学习

RAIL: Region-Aware Instructive Learning for Semi-Supervised Tooth Segmentation in CBCT

May 6, 2025

作者: Chuyu Zhao, Hao Huang, Jiashuo Guo, Ziyu Shen, Zhongwei Zhou, Jie Liu, Zekuan Yu

cs.AI

摘要

半監督學習已成為從CBCT掃描中進行三維牙齒分割的一種引人注目的方法，尤其是在標註數據極為有限的情況下。然而，現有方法仍面臨兩個持續存在的挑戰：在監督訓練過程中，對於結構模糊或錯誤標註區域的校正監督不足；以及由未標註數據上不可靠的偽標籤導致的性能下降。為解決這些問題，我們提出了區域感知指導學習（RAIL），這是一種雙組雙學生的半監督框架。每組包含兩個由共享教師網絡指導的學生模型。通過在兩組之間交替訓練，RAIL促進了組間知識轉移和協作式的區域感知指導，同時減少了對任何單一模型特性的過度擬合。具體而言，RAIL引入了兩種指導機制。分歧聚焦監督（DFS）控制器通過僅在學生輸出與真實標籤及最佳學生模型均存在差異的區域內指導預測，從而將監督集中於結構模糊或錯誤標註的區域，以此提升監督學習的效果。在無監督階段，置信度感知學習（CAL）調節器在模型確定性高的區域強化一致性，同時在訓練過程中降低低置信度預測的影響。這有助於防止模型學習不穩定的模式，並提高偽標籤的整體可靠性。在四個CBCT牙齒分割數據集上的廣泛實驗表明，RAIL在有限標註條件下超越了現有的最先進方法。我們的代碼將在https://github.com/Tournesol-Saturday/RAIL上公開。

English

Semi-supervised learning has become a compelling approach for 3D tooth segmentation from CBCT scans, where labeled data is minimal. However, existing methods still face two persistent challenges: limited corrective supervision in structurally ambiguous or mislabeled regions during supervised training and performance degradation caused by unreliable pseudo-labels on unlabeled data. To address these problems, we propose Region-Aware Instructive Learning (RAIL), a dual-group dual-student, semi-supervised framework. Each group contains two student models guided by a shared teacher network. By alternating training between the two groups, RAIL promotes intergroup knowledge transfer and collaborative region-aware instruction while reducing overfitting to the characteristics of any single model. Specifically, RAIL introduces two instructive mechanisms. Disagreement-Focused Supervision (DFS) Controller improves supervised learning by instructing predictions only within areas where student outputs diverge from both ground truth and the best student, thereby concentrating supervision on structurally ambiguous or mislabeled areas. In the unsupervised phase, Confidence-Aware Learning (CAL) Modulator reinforces agreement in regions with high model certainty while reducing the effect of low-confidence predictions during training. This helps prevent our model from learning unstable patterns and improves the overall reliability of pseudo-labels. Extensive experiments on four CBCT tooth segmentation datasets show that RAIL surpasses state-of-the-art methods under limited annotation. Our code will be available at https://github.com/Tournesol-Saturday/RAIL.