SemiTooth: a Generalizable Semi-supervised Framework for Multi-Source Tooth Segmentation

Abstract

With the rapid advancement of artificial intelligence, intelligent dentistry for clinical diagnosis and treatment has become increasingly promising. As the primary clinical dentistry task, tooth structure segmentation for Cone-Beam Computed Tomography (CBCT) has made significant progress in recent years. However, challenges arise from the obtainment difficulty of full-annotated data, and the acquisition variability of multi-source data across different institutions, which have caused low-quality utilization, voxel-level inconsistency, and domain-specific disparity in CBCT slices. Thus, the rational and efficient utilization of multi-source and unlabeled data represents a pivotal problem. In this paper, we propose SemiTooth, a generalizable semi-supervised framework for multi-source tooth segmentation. Specifically, we first compile MS3Toothset, Multi-Source Semi-Supervised Tooth DataSet for clinical dental CBCT, which contains data from three sources with different-level annotations. Then, we design a multi-teacher and multi-student framework, i.e., SemiTooth, which promotes semi-supervised learning for multi-source data. SemiTooth employs distinct student networks that learn from unlabeled data with different sources, supervised by its respective teachers. Furthermore, a Stricter Weighted-Confidence Constraint is introduced for multiple teachers to improve the multi-source accuracy.Extensive experiments are conducted on MS3Toothset to verify the feasibility and superiority of the SemiTooth framework, which achieves SOTA performance on the semi-supervised and multi-source tooth segmentation scenario.

Figures (5)

• Figure 1: Labeled and unlabeled samples of our Multi-Source Semi-Supervised MS$^3$Toothset. The comparisons of numerical (Density, Intensity) and feature distribution (t-SNE) illustrate the inherent source gaps. (Please zoom in for details.)
• Figure 2: SemiTooth Framework. (a) SemiTooth employs three students to handle different subsets, while two teachers supervise the mixed/other students to stabilize learning and improve pseudo-labels. (b) Mean Teacher is a basic framework that relies on a single teacher–student pair and lacks cross-source guidance. (c) Co-training uses multiple students (TaskNets) with shared weights but no teacher supervision, which fails to provide stable pseudo-labels. By combining multi-student collaboration and teacher guidance, SemiTooth effectively exploits multi-source data and enhances cross-source generalization.
• Figure 3: The Stricter Weighted-Confidence Constraint. SWC adaptively emphasizes reliable regions while suppressing noisy predictions, ensuring cleaner consistency learning across multi-source CBCT data. (Please zoom in for details.)
• Figure 4: Comparison with different methods and Ablation Study on SemiTooth. (a-b: Dental Images, Ground Truth). (c-g: SemiTooth, Uni-HSSL, CMT, MLRPL, ASDA, UA-MT, MT, V-Net). (k-o: Exp 5, 4, 3, 2, 1). (Please Zoom in for Details.)
• Figure 5: Multi-source feature comparisons after SemiTooth.