虚拟肝脏手术中高精度软组织建模与交互式仿真技术研究

Liver cancer is the second cancer killer in China. It is a serious threat to the health of our citizens. At present the surgical treatment is the most effective way and the first choice of treating moderate and advanced liver cancer. Virtual simulation system of liver surgery has already gradually been applied to clinical diagnosis and assisted the surgeon in pre-planning and surgical result predication to reduce surgical errors and patient distress. However, current virtual simulation system of liver surgery still has the problem of low-precision geometric shape and deformable model of liver soft tissue which greatly limit the system playing its role in the clinical diagnosis. Therefore, the project is to study the fuzzy image area enhancement algorithm of mixed multiscale geometric analysis and segmentation algorithm of nonlinear adaptive level set based on established 3D medical image workstation to get the geometric shape model of liver soft issue. And on that basis, the project further studies finite element physical modeling method of good performance to accurately simulate the deformation of liver soft issue. To further improve the performance of virtual surgical system, Solving the finite element of high quality on GPU is studied to fully indicate geometric shape model of liver tissue organ and deformation process details and assist the surgeon in getting more information. The research result will provide the solid theory foundation and technical support.

肝癌是我国第二位癌症杀手，严重威胁着我国公民健康，而外科手术治疗目前仍然是治愈中晚期肝癌最有效的方式也是首选方式。目前虚拟肝脏手术仿真系统已开始逐步应用于临床诊断，辅助医生进行手术前期规划和手术结果预测，以减少手术失误及减轻病人痛苦。然而，现有虚拟肝脏手术系统仍存在肝脏软组织几何形态与形变模型精度不高的问题，极大限制了其在临床诊断中发挥应有的作用。为此，本项目拟在已搭建的3D医学影像工作站基础上，研究基于混合多尺度几何分析的图像模糊区域增强算法及基于非线性自适应水平集的分割算法，以获得肝脏软组织的几何形态模型，在此基础上，研究性能优良的有限元物理建模方法，高精度地仿真肝脏软组织的形变；为了进一步提高虚拟手术系统的性能，研究基于GPU的有限元求解算法，实现交互式仿真，以充分显示肝脏器官的几何形态模型及形变过程细节，辅助医生获得更多的信息。该研究成果将为虚拟肝脏手术仿真提供坚实的理论基础和技术

目前虚拟手术仿真系统已开始逐步应用于临床诊断，辅助医生进行手术前期规划和手术结果预测。然而，现有虚拟手术系统仍存在器官软组织几何形态与形变模型精度不高等问题，极大限制了其在临床诊断中发挥应有的作用。为此，本项目拟在已搭建的3D医学影像工作站基础上，研究基于混合多尺度几何分析的图像模糊区域增强算法及基于非线性自适应水平集的分割算法，以获得器官软组织的几何形态模型，在此基础上，研究性能优良的有限元物理建模方法，高精度地仿真软组织的形变；为了进一步提高虚拟手术系统的性能，研究基于GPU　的有限元求解算法，实现交互式仿真，以充分显示器官的几何形态模型及形变过程细节，辅助医生获得更多的信息。