颌面缺损结构光三维测量与螺旋CT扫描图像的配准  

The study of registration of structured-light 3D facial data with spiral CT images of the maxillofacial defect

作　　者：邱憬[1,2] 孙健[1] 叶铭[3] 熊耀阳[1] 顾晓宇[1] 张富强[1] 

机构地区：[1]上海交通大学医学院附属第九人民医院口腔修复科,上海市口腔医学重点实验室,上海市200011 [2]江苏省口腔医院口腔修复科,江苏省南京市210029 [3]上海交通大学机械与动力工程学院,上海市200240

出　　处：《中国组织工程研究与临床康复》2010年第39期7319-7322,共4页Journal of Clinical Rehabilitative Tissue Engineering Research

基　　金：上海市经济和信息化委员会专项基金(0801031)"颌面缺损赝复数字化诊疗系统的开发及远程医疗服务与教育体系的建立";国家自然科学基金青年项目(30901694)"基于多数字化信息采集的颌面部缺损赝复体系的研究"~~

摘　　要：背景:目前尚无一种数字化信息采集技术能完全满足颌面缺损赝复的需要。因此,如能将光学三维测量与CT扫描数据的配准应用于颌面缺损,将为该疾病的诊断、治疗计划和赝复提供足够的软硬组织形态信息。目的:通过结构光三维测量与螺旋CT扫描重建模型的配准,构建一个高象素、复合内部骨结构的三维虚拟颌面缺损模型,并评估配准的精确度。方法:对上海第九人民医院口腔颌面外科收治的1名大面积复杂颌面缺损患者,分别使用自主研发的结构光三维光学测量系统和螺旋CT进行颌面部扫描和信息采集。分别使用Geomagic studio与自主研发的CAD-FacePros软件重建三维模型。使用CAD-FacePros软件,先通过面部解剖标志进行初配定位,再经迭代近邻点算法进行精确配准,将结构光测量获得的面部三维模型与CT重建后的颌面软组织模型进行匹配。使用CAD-FacePros软件计算两个匹配后模型间的最近点间距,获得模型配准误差。结果与结论:通过配准构建了一个高象素、复合内部骨结构的三维虚拟颌面缺损模型。平均配准误差为0.5mm,颌面部大部分区域的配准误差在1.0mm以内,颊部配准误差稍大。提示通过结构光三维测量与螺旋CT扫描重建模型的配准,构建高象素、复合内部骨结构的三维虚拟颌面缺损模型是可行的。BACKGROUND：No single technique can provide sufficient information to reconstruct a 3D model of a facial defect for clinical prosthetic treatment. Therefore,if it is possible to superimpose 3D optical soft-tissue image onto 3D CT scan data of a facial defect,this would provide enough information for diagnosis,planning,and treatment with high-resolution soft-tissue surface over the skull structure. OBJECTIVE：To assess the feasibility of merging surface images obtained individually by spiral CT scan and the structured-light scanning to develop an accurate,as well as complete,3D virtual patient model of maxillofacial defect combined with internal bone structures for prosthetic treatment. METHODS：A patient with an extensive and complex maxillofacial defect was selected from Department of Prosthodontics,Ninth People＇s Hospital. A fast structured-light scanning of the face using TDOS system and a spiral CT scan of the head were performed respectively. 3D reconstructions were correspondingly accomplished by Geomagic studio and CAD-FacePros. Using CAD-FacePros,the structured-light surface was surperimposed on the reconstructed CT soft-tissue surface using initial positioning based on corresponding landmarks and fine registration completed by the iterative closest point algorithm （ICP）. Through the use of CAD-FacePros,the registration errors were measured by calculating the discrepancy between the transformed structured-light and CT soft-tissue surfaces. RESULTS AND CONCLUSION：Using the registration method,an accurate and complete 3D virtual patient model of maxillofacial defect with underlying skeletal structure was acquired. The average registration error was 0.5 mm. The errors were within 1.0 mm in most parts of the aligned surfaces and were relatively large around the cheeks. A structured-light surface and spiral CT images can be superimposed to create an accurate and complete 3D virtual patient model of maxillofacial defect with underlying skeletal structure using the registration method.

关 键 词：配准 三维 结构光 螺旋CT 颌面缺损 

分 类 号：R318[医药卫生—生物医学工程]