基于特征线分段技术的牙齿分割算法

发布时间：2018-03-01 01:21

本文关键词： 三维牙颌模型牙齿分割牙齿形状建模形态学融合区域　出处：《计算机应用》2017年03期 　论文类型：期刊论文

【摘要】：牙齿分割是计算机口腔正畸的重要技术,针对三维牙颌模型直接进行牙齿分割而不对齿间融合区域进行处理会存在精确度较差、缺失侧面形状的问题,以及现有牙齿形状建模方法交互多、效率低的问题,提出一种基于特征线分段技术的牙齿分割算法。根据曲率信息筛选特征区域并采用形态学算法提取牙列特征线;结合特征线分段和分支点匹配算法以及形态学膨胀操作实现齿间融合区域的自动识别;利用匹配的分支点对齿间孔洞搭桥修补,实现牙齿形状的自动恢复;提取齿间龈缘线,然后以所有龈缘线作为牙齿分割线分离出单颗牙齿。实验结果表明,该算法不仅能准确分离出具有侧面形状的单颗牙齿,而且避免了牙齿形状建模时的交互操作,而且与手动识别并删除齿间粘连区域、采用曲面能量约束方式重建齿间缺失曲面的方法相比,提高了牙齿分割效率60%~90%。
[Abstract]:Tooth segmentation is an important technology in computer orthodontics. The problem of tooth segmentation without processing the fusion region between teeth in 3D dental model will be poor, and the problem of missing side shape will exist. As well as the problem that the existing tooth shape modeling methods have many interactions and low efficiency, a tooth segmentation algorithm based on feature line segmentation technology is proposed. The feature regions are filtered according to curvature information and the dentition feature lines are extracted by morphological algorithm. Combining feature line segmentation and branch point matching algorithm and morphological expansion operation to realize automatic recognition of fusion region between teeth; using matching branch points to repair holes between teeth to achieve automatic restoration of tooth shape; extracting gingival margin line between teeth, Then all gingival margin lines are used to separate a single tooth. Experimental results show that the algorithm can not only accurately separate a single tooth with side shape, but also avoid the interactive operation of tooth shape modeling. Compared with the method of manually recognizing and removing the adhesion area between teeth and using the method of surface energy constraint to reconstruct the missing surface between teeth, the efficiency of tooth segmentation is improved by 60% 90%.
【作者单位】：广东工业大学机电工程学院;
【基金】：国家自然科学基金资助项目(51175092) 广东省自然科学基金资助项目(10151009001000036)~~
【分类号】：R783.5;TP391.7

【相似文献】