<p>牙齿分割是牙齿矫正、辅助诊断等任务的关键环节，也是虚拟牙齿正畸系统的基础组件部分。牙齿分割所用到的牙齿数据类型主要包括三维牙颌模型和牙科全景X光影像，前者的特点体现在人的牙颌外观多变、形状结构复杂，几何特征信息丰富；后者在于影像整体对比度低，牙齿的像素值接近周边区域，牙齿的边界轮廓难以与周围组织区分开。因此，本文根据两种牙齿数据自身的特点，研究了基于深度学习的牙齿分割算法。</p> <p>本文的主要研究工作以及创新点如下：</p> <p>（1）针对现有算法只考虑到点坐标信息处理而忽略了法向量信息的问题，以及对于点坐标信息与法向量信息的特征学习未做有效区分的问题，提出了一种基于多重几何特征学习的三维牙齿分割方法。首先，采用空间变化网络作为网络输入中牙颌模型几何特征信息的校准对齐模块。然后，设计了一种多重几何特征学习模块，将牙颌模型中每个三角网格的中心点坐标和法向量编码增强，既保留了不同网格之间几何特征的差异，又使得不同属性的几何特征相互补充学习。最后，对于局部到全局的融合特征，采用多层感知器和高效通道注意力逐层串联的方式完成特征降维和通道优化，进一步改善三维牙颌模型的分割结果。实验结果表明，该方法取得了更高的牙齿分割精度，同时保持了较高的牙齿分割效率，无需繁琐复杂的人机交互，整体分割流程更加自动高效。</p> <p>（2）针对现有算法中多尺寸目标对象的提取能力弱的问题，提出了一种基于多特征坐标位置学习的牙科全景X光影像分割方法。首先，设计了一种残差全维度卷积模块，可以同时学习小尺寸细节特征和大尺寸区域特征，提高了网络的特征提取和恢复能力。此外，在特征编码网络和特征解码网络之间，设计了一种双流坐标注意力模块，进一步处理特征编码网络中每层的输出特征信息，整合并完成特征信息的丰富和优化。实验结果表明，该方法进一步提高了牙齿分割的准确度，牙齿的分割轮廓更加平滑，锯齿感明显减轻。</p> <p>综上所述，本文基于对牙齿数据特点的深入探究，提出了具有更高适应性和更优分割效果的牙齿分割方法，为提高虚拟牙齿正畸系统的自动化、智能化提供了重要的技术支撑。</p>

<p>Tooth segmentation is critical for orthodontics, auxiliary diagnostics, and serves as a fundamental component of virtual orthodontic systems. The types of dental data used mainly consist of 3D dental models and dental panoramic X-ray images. The 3D dental models are characterized by the variability of human dental appearance, the complexity of the shape and structure, and rich geometric feature information. Dental panoramic X-ray images, on the other hand, are characterized by low overall contrast, teeth pixel values that are similar to surrounding areas, and the difficulty in distinguishing tooth boundary contours from surrounding tissues. In light of these unique characteristics of the two types of dental data, this study proposes a tooth segmentation algorithm based on deep learning.</p> <p>(1) A method for three-dimensional tooth segmentation based on multiple geometric feature learning is proposed to address the problems of existing algorithms that only consider point coordinate information processing while neglecting normal vector information, and the ineffective discrimination of feature learning between point coordinate information and normal vector information. Firstly, a spatial transformer network is used as a calibration and alignment module for geometric feature information in the dental model as network input. Secondly, a multiple geometric feature learning module is designed to enhance the encoding of each triangle&#39;s center coordinates and normal vectors in the dental model, preserving the differences in geometric features between different meshes, while allowing geometric features of different attributes to mutually complement each other in learning. Finally, by using multi-layer perceptron and efficient channel attention methods in a layered and concatenated approach, local-to-global fusion features are realized and further improvements are made to the segmentation results of 3D dental models. Experimental results demonstrate that the proposed approach achieves higher precision for dental segmentation, while maintaining high efficiency without requiring cumbersome and complex human-computer interaction, thereby making the segmentation process more automatic and efficient.</p> <p>(2) A dental panoramic X-ray image segmentation method based on multi-feature coordinate position learning is proposed to address the weak ability of existing algorithms to extract multi-sized target objects. First, a residual omni-dimensional convolution module is designed to simultaneously learn small-scale detail features and large-scale regional features, thereby improving the network&#39;s feature extraction and restoration capabilities. In addition, a two-stream coordinate attention module is designed between the feature encoding network and the feature decoding network to further process the output feature information of each layer in the feature encoding network, integrating and optimizing the feature information. Experimental results demonstrate that the proposed method further improves the accuracy of tooth segmentation and produces smoother tooth segmentation contours with a noticeable reduction in jaggedness.</p> <p>In summary, this paper proposes a tooth segmentation method based on an in-depth exploration of the characteristics of tooth data, which is more adaptable and produces better segmentation results. This provides important technical support for the automation and intelligence of virtual tooth orthodontic systems.</p>

[1] 冯希平. 中国居民口腔健康状况—第四次中国口腔健康流行病学调查报告[C]//中华口腔医学会第十八次口腔预防医学学术年会论文汇编. 2018: 13⁃14.

[2] 孙乐海, 孙庆光. 三维数字化牙颌模型在口腔正畸学中的应用价值探讨[J]. 中国实用医药, 2021, 16(19): 137-139.

[3] Bianchi J, Mendonca G, Gillot M, et al. Three-Dimensional Digital Applications for Implant Space Planning in Orthodontics: A Narrative Review[J]. Journal of the World Federation of Orthodontists, 2022, 11(6): 207-215.

[4] Yang Y, Cui Z, Li C, et al. ToothInpaintor: Tooth Inpainting from Partial 3D Dental Model and 2D Panoramic Image[J]. arXiv preprint arXiv, 2022, 2211(15502): 1-10.

[5] Kondo Toshiaki, Ong S H, Foong Kelvin W C. Tooth Segmentation of Dental Study Models using Range Images[J]. IEEE Transactions on Medical Imaging, 2004, 23(3): 350-362.

[6] 郝国栋, 程筱胜, 戴宁, 俞青. 基于形态学的牙齿模型交互分割[J]. 中国制造业信息化, 2008(01): 36-39.

[7] Yuan T, Liao W, Dai N, et al. Single-Tooth Modeling for 3D Dental Model[J]. International Journal of Biomedical Imaging, 2010, 2010: 1-14.

[8] Kronfeld T, Brunner D, Brunnett G. Snake-Based Segmentation of Teeth from Virtual Dental Casts[J]. Computer-Aided Design and Applications, 2010, 7(2): 221-233.

[9] Wu K, Chen L, Li J, et al. Tooth Segmentation on Dental Meshes using Morphologic Skeleton[J]. Computers & Graphics, 2014, 38: 199-211.

[10] Zou B, Liu S, Liao S, et al. Interactive Tooth Partition of Dental Mesh Base on Tooth-Target Harmonic field[J]. Computers in Biology and Medicine, 2015, 56: 132-144.

[11] Mangan A P, Whitaker R T. Partitioning 3D Surface Meshes using Watershed Segmentation[J]. IEEE Transactions on Visualization and Computer Graphics, 1999, 5(4): 308-321.

[12] Li Z, Ning X, Wang Z. A Fast Segmentation Method for STL Teeth Model[C]//IEEE/ICME International Conference on Complex Medical Engineering. 2007: 163-166.

[13] Kumar Y, Janardan R, Larson B, et al. Improved Segmentation of Teeth in Dental Models [J]. Computer-Aided Design and Applications, 2011, 8(2): 211-224.

[14] Qi C R, Su H, Mo K, et al. PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation[C]//IEEE Conference on Computer Vision and Pattern Recognition. 2017: 652-660.

[15] Qi C R, Yi L, Su H, et al. PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space[C]//International Conference on Neural Information Processing Systems, 2017: 5105-5114.

[16] Wang L, Huang Y, Hou Y, et al. Graph Attention Convolution for Point Cloud Semantic Segmentation[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2019: 10288-10297.

[17] Hu Q, Yang B, Xie L, et al. RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2020: 11105-11114.

[18] Xu X, Liu C, Zheng Y. 3D Tooth Segmentation and Labeling using Deep Convolutional Neural Networks[J]. IEEE Transactions on Visualization and Computer Graphics, 2018, 25(7): 2336-2348.

[19] 李占利, 孙志浩, 李洪安, 刘童鑫. 图卷积网络下牙齿种子点自动选取[J]. 中国图象图形学报, 2020, 25(07): 1481-1489.

[20] Verma N, Boyer E, Verbeek J. FeaStNet: Feature-Steered Graph Convolutions for 3D Shape Analysis[C]//IEEE Conference on Computer Vision and Pattern Recognition. 2018: 2598-2606.

[21] 田素坤, 戴宁, 袁福来, 张贝, 俞青, 程筱胜. 多级层次三维卷积神经网络的牙颌模型分割与识别技术[J]. 计算机辅助设计与图形学学报, 2020, 32(08): 1218-1227.

[22] 张雅玲, 于泽宽, 何炳蔚. 基于GCNN的CBCT模拟口扫点云数据牙齿分割算法[J]. 计算机辅助设计与图形学学报, 2020, 32(07): 1162-1170.

[23] 王志江, 秦品乐, 柴锐, 武峰, 程一彤, 史玥. 基于深度学习的牙齿嵌塞自动判别方法[J]. 计算机工程, 2022, 48(04): 307-313.

[24] Thomas H, Qi C R, Deschaud J E, et al. KPConv: Flexible and Deformable Convolution for Point Clouds[C]//IEEE/CVF International Conference on Computer Vision. 2019: 6410-6419.

[25] Lian C, Wang L, Wu T H, et al. Deep Multi-Scale Mesh Feature Learning for Automated Labeling of Raw Dental Surfaces From 3D Intraoral Scanners[J]. IEEE Transactions on Medical Imaging, 2020, 39(7): 2440-2450.

[26] Zanjani F G, Pourtaherian A, Zinger S, et al. Mask-MCNet: Tooth Instance Segmentation in 3D Point Clouds of Intra-Oral Scans[J]. Neurocomputing, 2021, 453: 286-298.

[27] Li Y, Bu R, Sun M, et al. PointCNN: Convolution On X-Transformed Points[C]//Advances in Neural Information Processing Systems. 2018: 820-830.

[28] Cui Z, Li C, Chen N, et al. TSegNet: An Efficient and Accurate Tooth Segmentation Network on 3D Dental Model[J]. Medical Image Analysis, 2021, 69: 1-12.

[29] Zhang L, Zhao Y, Meng D, et al. TSGCNet: Discriminative Geometric Feature Learning With Two-Stream Graph Convolutional Network for 3D Dental Model Segmentation[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2021: 6699-6708.

[30] Radhiyah A, Harsono T, Sigit R. Comparison Study of Gaussian and Histogram Equalization Filter on Dental Radiograph Segmentation for Labelling Dental Radiograph [C]//International Conference on Knowledge Creation and Intelligent Computing. 2016: 253-258.

[31] Li H, Sun G, Sun H, et al. Watershed Algorithm Based on Morphology for Dental X-Ray Images Segmentation[C]//International Conference on Signal Processing. 2012: 877-880.

[32] Tikhe S V, Naik A M, Bhide S D, et al. Algorithm to Identify Enamel Caries and Interproximal Caries using Dental Digital Radiographs[C]//International Conference on Advanced Computing. 2016: 225-228.

[33] Mao J, Wang K, Hu Y, et al. GrabCut Algorithm for Dental X-Ray Images Based on Full Threshold Segmentation[J]. IET Image Processing, 2018, 12(12): 2330-2335.

[34] Said E H, Nassar D E M, Fahmy G, et al. Teeth Segmentation in Digitized Dental X-ray Films using Mathematical Morphology[J]. IEEE Transactions on Information Forensics and Security, 2006, 1(2): 178-189.

[35] Alsmadi M K. A Hybrid Fuzzy C-Means and Neutrosophic for Jaw Lesions Segmentation[J]. Ain Shams Engineering Journal, 2018, 9(4): 697-706.

[36] Tuan T M. A Cooperative Semi-Supervised Fuzzy Clustering Framework for Dental X-Ray Image Segmentation[J]. Expert Systems with Applications, 2016, 46: 380-393.

[37] Trivedi D N, Modi C K. Dental Contour Extraction using ISEF Algorithm for Human Identification[C]//International Conference on Electronics Computer Technology. 2011: 6-10.

[38] Niroshika U A A, Meegama R G N, Fernando T G I. Active Contour Model to Extract Boundaries of Teeth in Dental X-Ray Images[C]//International Conference on Computer Science & Education. 2013: 396-401.

[39] Modi C K, Desai N P. A Simple and Novel Algorithm for Automatic Selection of ROI for Dental Radiograph Segmentation[C]//Canadian Conference on Electrical and Computer Engineering. 2011: 000504-000507.

[40] 於猛, 刘向荣, 黄晓阳, 陈迎潮. 改进区域生长法及其在冠状动脉造影图中的应用[J]. 厦门大学学报(自然科学版), 2013, 52(01): 38-42.

[41] 章斌, 卢洪义, 刘舜, 桑豆豆, 杨禹成. 发动机部件CT图像特征提取与区域生长算法[J]. 兵工学报, 2022: 1-10.

[42] Long J, Shelhamer E, Darrell T. Fully Convolutional Networks for Semantic Segmentation[C]//IEEE Conference on Computer Vision and Pattern Recognition. 2015: 3431-3440.

[43] Ronneberger O, Fischer P, Brox T. U-Net: Convolutional Networks for Biomedical Image Segmentation[C]//Medical Image Computing and Computer-Assisted Intervention. 2015: 234-241.

[44] He K, Zhang X, Ren S, et al. Deep Residual Learning for Image Recognition[C]//IEEE Conference on Computer Vision and Pattern Recognition. 2016: 770-778.

[45] Huang G, Liu Z, Van Der Maaten L, et al. Densely Connected Convolutional Networks[C]//IEEE Conference on Computer Vision and Pattern Recognition. 2017: 4700-4708.

[46] Zhou Z, Siddiquee M M R, Tajbakhsh N, et al. UNet++: A Nested U-Net Architecture for Medical Image Segmentation[C]//Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support. 2018: 3-11.

[47] Liu S, Li Y, Zhou J, et al. Segmenting Nailfold Capillaries using an Improved U-Net Network[J]. Microvascular Research, 2020, 130: 1-9.

[48] Zhuang J. LadderNet: Multi-Path Networks Based on U-Net for Medical Image Segmentation[J]. arXiv preprint arXiv, 2018, 1810(07810): 1-4.

[49] Shuvo M B, Ahommed R, Reza S, et al. CNL-UNet: A Novel Lightweight Deep Learning Architecture for Multimodal Biomedical Image Segmentation with False Output Suppression[J]. Biomedical Signal Processing and Control, 2021, 70: 1-14.

[50] Rao Y, Nartey O T, Zeng S, et al. LeFUNet: UNet with Learnable Feature Connections for Teeth Identification and Segmentation in Dental Panoramic X-ray Images[C]//IEEE International Conference on Bioinformatics and Biomedicine. IEEE, 2022: 2110-2118.

[51] Moran M, Faria M, Giraldi G, et al. Classification of Approximal Caries in Bitewing Radiographs using Convolutional Neural Networks[J]. Sensors, 2021, 21(15): 1-12.

[52] Zhao H., Shi J., Qi X., et al. Pyramid Scene Parsing Network[C]//IEEE Conference on Computer Vision and Pattern Recognition. 2017: 2881-2890.

[53] Gu Z, Cheng J, Fu H, et al. CE-Net: Context Encoder Network for 2D Medical Image Segmentation[J]. IEEE Transactions on Medical Imaging, 2019, 38(10): 2281-2292.

[54] Wang G, Liu H, Yi X, et al. ARMS Net: Overlapping Chromosome Segmentation Based on Adaptive Receptive Field Multi-Scale Network[J]. Biomedical Signal Processing and Control, 2021, 68: 1-9.

[55] Wang D, Hu G, Lyu C. Multi-Path Connected Network for Medical Image Segmentation[J]. Journal of Visual Communication and Image Representation, 2020, 71: 1-11.

[56] Chen Q, Zhao Y, Liu Y, et al. MSLPNet: Multi-Scale Location Perception Network for Dental Panoramic X-Ray Image Segmentation[J]. Neural Comput & Applic, 2021, 33: 10277-10291.

[57] Kose K, Bozkurt A, Alessi-Fox C, et al. Segmentation of Cellular Patterns in Confocal Images of Melanocytic Lesions in Vivo Via a Multiscale Encoder-Decoder Network (MED-Net)[J]. Medical Image Analysis, 2021, 67: 1-12.

[58] Vaswani A, Shazeer N, Parmar N, et al. Attention is All You Need[C]//International Conference on Neural Information Processing Systems. 2017: 6000-6010.

[59] Oktay O, Schlemper J, Folgoc L L, et al. Attention U-Net: Learning Where to Look for the Pancreas[J]. arXiv preprint arXiv, 2018, 1804(03999): 1-10.

[60] 张淑军, 彭中, 李辉. SAU-Net：基于U-Net和自注意力机制的医学图像分割方法[J]. 电子学报, 2022, 50(10): 2433-2442.

[61] Li Y, Wang S, Wang J, et al. GT U-Net: A U-Net Like Group Transformer Network for Tooth Root Segmentation[C]//Machine Learning in Medical Imaging. 2021: 386-395.

[62] Heo B, Yun S, Han D, et al. Rethinking Spatial Dimensions of Vision Transformers[C]//IEEE/CVF International Conference on Computer Vision. 2021: 11936-11945.

[63] Gu R, Wang G, Song T, et al. CA-Net: Comprehensive Attention Convolutional Neural Networks for Explainable Medical Image Segmentation[J]. IEEE Transactions on Medical Imaging, 2020, 40(2): 699-711.

[64] Kaya M C, Akar G B. Dental X-Ray Image Segmentation using Octave Convolution Neural Network[C]//Signal Processing and Communications Applications Conference (SIU). IEEE, 2020: 1-4.

[65] Chen Y, Fan H, Xu B, et al. Drop an Octave: Reducing Spatial Redundancy in Convolutional Neural Networks with Octave Convolution[C]//IEEE/CVF international conference on computer vision. 2019: 3435-3444.

[66] Sheng C, Wang L, Huang Z, et al. Transformer-Based Deep Learning Network for Tooth Segmentation on Panoramic Radiographs[J]. Journal of Systems Science and Complexity, 2023, 36(1): 257-272.

[67] Cao H, Wang Y, Chen J, et al. Swin-Unet: Unet-Like Pure Transformer for Medical Image Segmentation[C]//European Conference on Computer Vision. 2023: 205-218.

[68] Kong Z, Xiong F, Zhang C, et al. Automated Maxillofacial Segmentation in Panoramic Dental X-Ray Images using an Efficient Encoder-Decoder Network[J]. IEEE Access, 2020, 8: 207822-207833.

[69] Zhao Y, Li P, Gao C, et al. TSASNet: Tooth Segmentation on Dental Panoramic X-Ray Images by Two-Stage Attention Segmentation Network[J]. Knowledge-Based Systems, 2020, 206: 1-10.

[70] Valanarasu J M J, Patel V M. UNeXt: MLP-Based Rapid Medical Image Segmentation Network[C]//Medical Image Computing and Computer Assisted Intervention. 2022: 23-33.

[71] Wang Q, Wu B, Zhu P, et al. ECA-Net: Efficient Channel Attention for Deep Convolutional Neural Networks[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2020: 11531-11539.

[72] Jaderberg M, Simonyan K, Zisserman A. Spatial Transformer Networks[C]//International Conference on Neural Information Processing Systems. 2015: 2017-2025.

[73] 关柏良, 周凡, 林淑金, 罗笑南. 面向三维模型分割的边界感知点云神经网络[J]. 计算机辅助设计与图形学学报, 2020, 32(01): 147-155.

[74] Sudre C H, Li W, Vercauteren T, et al. Generalised Dice Overlap as a Deep Learning Loss Function for Highly Unbalanced Segmentations[C]//Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support. 2017: 240-248.

[75] Smith L N. Cyclical Learning Rates for Training Neural Networks[C]//IEEE Winter Conference on Applications of Computer Vision. 2017: 464-472.

[76] Li C, Zhou A, Yao A. Omni-Dimensional Dynamic Convolution[C]//International Conference on Learning Representations. 2022: 1-20.

[77] Yang B, Bender G, Le Q V, et al. CondConv: Conditionally Parameterized Convolutions for Efficient Inference[C]//International Conference on Neural Information Processing Systems. 2019: 1307-1318.

[78] Chen Y, Dai X, Liu M, et al. Dynamic Convolution: Attention over Convolution Kernels[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2020: 11030-11039.

[79] Hu J, Shen L, Sun G. Squeeze-and-Excitation Networks[C]//IEEE Conference on Computer Vision and Pattern Recognition. 2018: 7132-7141.

[80] Woo S, Park J, Lee J Y, et al. CBAM: Convolutional Block Attention Module[C]//European Conference on Computer Vision. 2018: 3-19.

[81] Hou Q, Zhou D, Feng J. Coordinate Attention for Efficient Mobile Network Design[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2021: 13713-13722.

[82] Abdi A H, Kasaei S, Mehdizadeh M. Automatic Segmentation of Mandible in Panoramic X-Ray[J]. Journal of Medical Imaging, 2015, 2(4): 1-8.