基于区域能量函数的快速CBCT图像牙齿分割算法  

A Regional Energy Function Based Approach for Fast Tooth Segmentation from CBCT Images

作　　者：钱驾宏 孙夏 汤德衍 陶煜波[1,2] 林军[3] 林海[1,2] Qian Jiahong;Sun Xia;Tang Deyan;Tao Yubo;Lin Jun;Lin Hai(State Key Laboratory of CAD ＆ CG, Zhejiang University, Hangzhou 310058;Innovation Center for Minimally Invasive Technique and Device, Zhejiang University, Hangzhou 310020;The First Affiliated Hospital, Zhejiang University, Hangzhou 310006)

机构地区：[1]浙江大学CAD&CG国家重点实验室,杭州310058 [2]浙江大学微创技术和器械临床医学创新中心,杭州310020 [3]浙江大学附属第一医院,杭州310006

出　　处：《计算机辅助设计与图形学学报》2018年第6期975-983,共9页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金(61472354,61672452); NSFC-广东省人民政府大数据科学研究中心项目(U1611263); 中央高校基本科研业务费专项资金

摘　　要：从锥形束面计算机断层扫描图像中分割出三维牙齿模型对口腔正畸治疗和研究工作十分重要.针对邻接牙、牙周组织模糊等问题,提出一种基于区域能量函数的快速牙齿分割算法.首先通过牙髓种子点和梯度模,筛选拟合得到一条粗略的初始分割曲线;然后根据灰度分布特征确定曲线变化的搜索路径和变化区域,在此区域内构造了一个基于灰度值、梯度模和边界平滑性的区域能量函数,优化初始分割曲线得到精确的分割曲线;再将上一层的最终分割曲线作为下一层的初始分割曲线,逐层优化得到所有分割曲线;最后使用Marching Cubes算法重建得到完整牙齿网格模型.采用CBCT数据集进行测试,实验结果表明,该算法能够有效且快速地解决模糊问题,得到满足正畸治疗需要的牙齿模型.Segmenting 3 D teeth models from cone beam computed tomography images is significantly important for orthodontics. In this paper, we propose a new approach based on a regional energy function to solve the fuzziness problem of adjacent teeth crowns and root-jaw issues. The initial curve is first calculated based on the endodontic seeds and gradients. Then, the moving range of the initial curve is restricted by the intensity distribution. Within this moving range, we propose a regional energy function based on intensity, gradient and smoothness. By optimizing the function, the final curve of the slice can be generated, and is applied as the initial curve of next slice. The final tooth model can be reconstructed by Marching Cubes algorithm. The experimental results on CBCT datasets demonstrate that our method is accurate and efficient for the fuzziness problem in orthodontic treatment.

关 键 词：CBCT图像 牙齿分割 区域能量函数 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]