TCATSeg: A Tooth Center-Wise Attention Network for 3D Dental Model Semantic Segmentation

Qiang He; Wentian Qu; Jiajia Dai; Changsong Lei; Shaofeng Wang; Feifei Zuo; Yajie Wang; Yaqian Liang

cs.CVMar 17, 2026

TCATSeg: A Tooth Center-Wise Attention Network for 3D Dental Model Semantic Segmentation

Authors: Qiang He, Wentian Qu, Jiajia Dai, Changsong Lei, Shaofeng Wang, Feifei Zuo, Yajie Wang, Yaqian Liang, +4 more

Organizations: Institute of Software, Chinese Academy of Sciences · University of Chinese Academy of Sciences · Department of Computer Science and Technology, Tsinghua University · Beijing Stomatological Hospital, Capital Medical University · LargeV Instrument Corporation, Ltd.

Abstract

Accurate semantic segmentation of 3D dental models is essential for digital dentistry applications such as orthodontics and dental implants. However, due to complex tooth arrangements and similarities in shape among adjacent teeth, existing methods struggle with accurate segmentation, because they often focus on local geometry while neglecting global contextual information. To address this, we propose TCATSeg, a novel framework that combines local geometric features with global semantic context. We introduce a set of sparse yet physically meaningful superpoints to capture global semantic relationships and enhance segmentation accuracy. Additionally, we present a new dataset of 400 dental models, including pre-orthodontic samples, to evaluate the generalization of our method. Extensive experiments demonstrate that TCATSeg outperforms state-of-the-art approaches.

Figures & tables

Figure 1: SOTA method [ 22 ] struggles on our TeethWild dataset.

Figure 2: Pipeline of TCATSeg. The encoder introduces two novel components: DPDA, which enables superpoints (TCP) to capture semantic context, and SGDA, which integrates local geometric features with global semantic context.

Figure 3: Tooth segmentation results on Teeth3DS and TeethWild, with errors highlighted by red dashed lines.

D.	Method	TIR $\uparrow$	SEN $\uparrow$	DSC $\uparrow$	PPV $\uparrow$	OA $\uparrow$
Teeth3DS	TSGCNet	95.17	92.32	89.64	91.26	94.67
	DBGANet	95.91	94.30	89.82	90.65	95.54
	DGCNN	95.77	91.73	89.94	92.23	94.35
	PointNet++	95.31	93.17	90.87	92.40	95.04
	PT	95.80	95.29	93.51	94.58	96.30
	SPoTr	95.92	95.13	93.45	94.49	96.15

Table 1: Segmentation results on Teeth3DS and TeethWild.

Method	TLA $\uparrow$	TSA $\uparrow$	TIR $\uparrow$	Score $\uparrow$
CGIP	96.58	98.59	91.00	95.39
FiboSeg	99.24	92.93	92.23	94.80
IGIP	92.44	97.50	92.89	94.27
TeethSeg	91.84	96.78	85.38	91.33
OS	78.45	96.93	89.40	88.26
Chompers	62.42	88.86	87.95	79.74

Table 2: The results of 3DTeethSeg’22 Challenge [ 2 ] .

Figure 4: The comparison of superpoints in SpoTr and TCATSeg. Red points correspond to the superpoints.

Explore similar work

B-Spline Embedded Structure Learning for 3D Tooth Segmentation
Aug 18, 2026Xianghan Wei, Jianwen Lou, Zhiguo Lu +23D Medical Image Segmentation 3D Semantic Segmentation
School of Software Technology, Zhejiang University · School of Computer and Computing Science, Hangzhou City University · The Affiliated Stomatology Hospital, Zhejiang University School of Medicine
TSegAgent: Zero-Shot Tooth Segmentation via Geometry-Aware Vision-Language Agents
Mar 20, 2026Shaojie Zhuang, Lu Yin, Guangshun Wei +3Medical Imaging Medical Image Segmentation
Shandong University, Jinan 250010, China · University of Surrey, Guildford, Surrey, United Kingdom · King’s College London, London, United Kingdom
Innovative Tooth Segmentation Using Hierarchical Features and Bidirectional Sequence Modeling
Feb 25, 2026Xinxin Zhao, Jian Jiang, Yan Tian +5Medical Image Segmentation
School of Computer Science and Technology, Zhejiang Gongshang University, 310018, Hangzhou, China · School of Computer and Cyber Sciences, Communication University of China, 100024, Beijing, China · Center of Big Data, China Digital Culture Group Co., Ltd, 100176, Beijing, China +6

Aug 18, 2026J/K move · Enter open · S save

B-Spline Embedded Structure Learning for 3D Tooth Segmentation

Xianghan Wei, Jianwen Lou, Zhiguo Lu, Hairong Jin+1

School of Software Technology, Zhejiang University · School of Computer and Computing Science, Hangzhou City University · The Affiliated Stomatology Hospital, Zhejiang University School of Medicine

3D Medical Image Segmentation 3D Semantic Segmentation 3D Representation Learning

Open paper AskPDF arXiv

Accurate 3D tooth segmentation forms the cornerstone of digital dentistry, yet it remains a formidable challenge due to the inherent intricacy of real-world dentitions, such as crowding, misaligned teeth and high morphological similarity between adjacent teeth. To resolve this, we present B-Spline Embedded Structure Learning, a novel framework that distills the inherent sequential arrangement of teeth into a continuous structural constraint to regularize representation space. Our approach parameterizes the global dental topology by fitting a parametric B-spline trajectory to tooth centers, assigning each point a continuous structural embedding that forces the shared backbone to capture global arch organization. To fully exploit these embedded priors, we introduce a Structure-Aware Dynamic Classifier (SADC) to substitute rigid static templates with adaptive, case-calibrated decision boundaries. SADC regularizes dynamic prototype pooling via a localized Gaussian proximity gate and contextually co-evolves them through an attention block modeling spatial relations and bilateral symmetries across teeth. Extensive evaluations on the 3DTeethSeg22 benchmark demonstrate that our method establishes a new state-of-the-art accuracy with exceptional structural robustness and efficiency in computational overhead, markedly enhancing the model's capacity to handle complex dental configurations.

Abstract

Explore similar work

B-Spline Embedded Structure Learning for 3D Tooth Segmentation

TSegAgent: Zero-Shot Tooth Segmentation via Geometry-Aware Vision-Language Agents

Innovative Tooth Segmentation Using Hierarchical Features and Bidirectional Sequence Modeling