利用数字几何技术重建个性化骨骼模型

发布时间：2019-08-13 11:57

【摘要】：背景:由于医学CT体数据存在各向异性的特点,导致CT序列图像重建网格模型时产生阶梯表面,从而影响后续的医学诊断。目的:利用数字几何处理技术重建个性化骨骼模型。方法:首先基于互信息的图像配准算法对骨骼CT序列图像进行配准,接着使用图像分割提取骨骼轮廓集并转化为三维点云,然后使用高斯加权的主成分分析方法估算点云法向量并对点云进行三边滤波去噪,最后对点云进行自适应圆球覆盖及网格化处理,完成个性化骨骼模型重建。结果与结论:文章所提的方法可以生成光顺的个性化骨骼表面网格模型,所形成的三角网格形状规则且自适应分布,可以为计算机辅助制造、有限元分析及3D打印提供准确的三维模型。
[Abstract]:Background: due to the anisotropy of medical CT volume data, the ladder surface of CT sequence image reconstruction grid model is produced, which affects the subsequent medical diagnosis. Objective: to reconstruct personalized bone model with digital geometry processing technology. Methods: firstly, bone CT sequence images were registered based on mutual information image registration algorithm, then the bone outline set was extracted and transformed into three-dimensional point cloud by image segmentation, then the point cloud normal vector was estimated by Gaussian weighted principal component analysis method and the point cloud was filtered and de-noised by trilateral filtering. Finally, the point cloud was covered and gridded adaptively to complete the reconstruction of personalized bone model. Results and conclusion: the method proposed in this paper can generate a smooth personalized bone surface grid model, and the triangular grid is regular and adaptive distribution, which can provide an accurate three-dimensional model for computer-aided manufacturing, finite element analysis and 3D printing.
【作者单位】：淮阴工学院江苏省先进制造重点实验室;淮安市淮阴医院骨科;
【基金】：江苏省高校自然科学研究面上项目(14KJB460002) 淮安市科技计划项目(HAG2015032) 江苏省先进制造重点实验室开放课题(HGDML-1204)~~
【分类号】：R816.8;R68

【相似文献】