基于面下1/3软组织全貌虚拟预测的全口义齿数字化设计

According to the latest oral epidemiological survey data, there are about 10 million edentulous patients requiring to be restored with complete denture in China. The prediction of fullness and aesthetics of the complete denture is an important base for denture design and evaluation. Until now, no CAD system for complete denture has this function. Therefore, this research is intended to study the virtual prediction of face appearance changes after the complete denture being inserted into edentulous patients' mouth. The sample consists of 50 edentulous patients with five kinds of jaw arch. Each patient will wear five types of anterior teeth protrusions complete denture and then accept facial three-dimensional scanning separately. The Principal Component Analysis will be used for analyzing main geometric factors of face appearance such as the nasolabial fold, and the standard topology template of lower 1/3 soft tissue geometric full view of edentulous patients will be established. Meanwhile, based on Multiple Linear Regression and the Laplace deformation function, the control algorithm of the facial soft tissue deformation will be researched and the mathematical model for predicting the relationship between artificial dentition and lower 1/3 appropriate fullness will be established. Finally, these methods will be recombined with the complete denture CAD/RP system developed by our research group. Hand-crafted wax rim and denture trying-in processes will be replaced by digital and automatic technologies, and subjective qualitative way will be replaced by objective qualitative way, the clinical operation procedure of complete denture will be simplified, the difficulty of the operation will be reduced and the restoration effect will be improved.

据最新口腔流调数据，我国现有约1000万无牙颌患者需用全口义齿修复，戴入义齿后患者面容丰满度及美学的预测是临床修复设计和效果评价的重要依据。国内外目前正在研发的全口义齿CAD系统均不具备此功能。因此，本项目拟针对无牙颌患者戴牙前后面容变化的虚拟预测展开研究。样本由5种颌弓型，共50例无牙颌患者组成。每个患者分别戴入5种不同前牙突度的义齿后进行面部三维扫描。对获取的数据如鼻唇沟等面容主要形态构成成份进行分析，建立无牙颌面下1/3软组织几何全貌标准拓扑模板；同时，基于多元线性回归和拉普拉斯曲面变形函数，研究面部软组织变形控制算法，建立能预测人工牙列与面下1/3适宜丰满度关系的数学模型。最终与本项目组已自行开发的全口义齿CAD/RP辅助设计制作系统进行重组，用数字化、自动化技术取代手工制作蜡颌堤和试排牙过程，用客观定量替代主观定性，简化全口义齿修复的临床操作过程，降低操作难度，提高修复效果。

据最新口腔流调数据，我国现有约1000万无牙颌患者需用全口义齿修复，戴入义齿后患者面容丰满度及美学的预测是临床修复设计和效果评价的重要依据。国内外目前正在研发的全口义齿CAD系统均不具备此功能。因此，本项目拟针对无牙颌患者戴牙前后面容变化的虚拟预测展开研究。样本由5种颌弓型，共50例无牙颌患者组成。每个患者分别戴入5种不同前牙突度的义齿后进行面部三维扫描。对获取的数据如鼻唇沟等面容主要形态构成成份进行分析，建立无牙颌面下1/3软组织几何全貌标准拓扑模板；同时，基于多元线性回归和拉普拉斯曲面变形函数，研究面部软组织变形控制算法，建立能预测人工牙列与面下1/3适宜丰满度关系的数学模型。本项目的研究综合口腔修复学、口腔解剖生理学、医学统计学、计算机图形学、设计方法学等多个学科内容，采用主成分分析、多元线性回归和STL细分曲面变形等方法建立无牙颌面下1/3软组织几何全貌拓扑特征模板和识别典型区域个性化参数，建立全口义齿修复后适宜面下1/3容貌虚拟预测的基础数学模型。进一步，考虑到人体皮肤组织是非均匀材料，具有非线性、生物组织结构的特殊性以及物理机械性能的特异性，而BP神经网络能够表达复杂的非线性系统，于是利用BP神经网络获取无牙颌模型的特征模板与弹性变形之间的关系，构建面部弹性变形预测模型。探索出一条基于面容虚拟预测的全口义齿数字化的快速设计新思路。最终与本项目组已自行开发的全口义齿CAD/RP辅助设计制作系统进行重组，用数字化、自动化技术取代手工制作蜡颌堤和试排牙过程，用客观定量替代主观定性，简化全口义齿修复的临床操作过程，降低操作难度，提高修复效果。