基于图论分割的肺部CT图像的三维重建

为了得到精准的人体肺部CT图像的分割结果,采用改进的最小生成树法对人体肺部CT图像进行分割,再采用面绘制中的Marching Cubes(MC)算法进行三维重建,实现肺部的三维立体显示.通过实验仿真,验证了改进最小生成树算法的快速有效性,并将该算法与基于阈值分割的三维重建仿真效果进行对比.结果表明,改进后的算法能有效提高肺部CT图像三维重建的效率和完整度,在保证了快速三维重建的同时,三维重建的效果更佳,将为医生的医疗诊断提供有力的判断依据.

Three-dimensional reconstruction of lung CT images based on graph theory segmentation

In order to obtain the accurate segmentation results of human lung CT images, the improved minimum spanning tree method was used to segment the human lung CT images. The Marching Cubes （MC） algorithm in surface rendering was used to perform the three dimensional （3D） reconstruction, which realized the 3D stereoscopic display for human lung. The speediness and effectiveness of improved minimum spanning tree algorithm were verified with simulation experiments. In addition, the algorithm was compared with the simulated effect of 3D reconstruction based on threshold segmentation. The results show that the improved algorithm can enhance the efficiency and integrity of 3D reconstruction of lung CT images. The better effect of 3D reconstruction can be obtained when the fast 3D reconstruction is ensured at the same time. The study can provide a strong judgment basis for the medical diagnosis of doctors.