SPECT医学图像重建及三维可视化研究

Study on SPECT Medical Image Reconstruction and 3D Visualization

针对医学图像的建模与仿真分析需要解决的几何模型精度以及物理模型的正确性问题。本文以人体牙齿CT图像为例实现建模过程,在牙齿分割方面提出了改进。通过读取DICOM图像,对断层图像进行裁剪和分割等预处理,使用移动立方体(MC)算法进行表面重构,同时使用二次误差度量(QEM)方法消减和平滑三角面片,使用德洛奈(Delaunay)方法由面网格推出体网格模型。最后根据目标组织选择相应的物理模型采用有限元(FEM)方法做分析,使用Open GL显示等效应力的受力云图。分析结果显示,本文所进行的SPECT医学图像重建及三维可视化研究结果与理论相符,通过SPECT图像的重建,更加准确的了解到人体组织器官的功能和代谢情况,提高了医护人员对疾病诊断的准确率。

Aiming at solving the problem of medical image accuracy of geometry model and physical model in the analysis of modeling and simulation. This paper took teeth CT image as an example to realize human teeth modeling process, and put forward to improve the tooth segmentation. By reading the DICOM image, the tomographic images were cropped and segmenting, and used the marching cubes (MC) algorithm for surface reconstruction. At the same time, we used two quadric error metric (QEM) method to reduce and smooth triangle piece, and used the method of Delaunay to deduce the surface mesh to the solid mesh. Finally, we analyzed the target tissue with the finite element method (FEM) according to the choice of the corresponding physical model, and used the OpenGL to display the stress cloud atlas of equivalent stress. The results showed that the reconstruction of medical image SPECT and 3D visualization were consistent with the theory. Doctors could understand the human tissue and organ more accurately through the reconstruction of SPECT images, and the correct rate of diagnosis on disease had been improved.