基于元球和无网格模型混合方法的可交互软组织建模及其在神经外科手术模拟中的应用

随着虚拟现实技术在医学领域的不断应用，虚拟手术、手术规划等越来越受到临床的关注,尤其是在神经外科领域。虚拟手术研究的核心是人体组织的建模及手术操作的模拟。元球模型作为一种特殊的隐式曲面造型技术，可以有效描述光滑、圆润的曲面。人脑皮层有着众多圆润的沟回和褶皱，这些复杂的几何结构如果用元球来表达和简化，在处理碰撞检测和变形将具有天然的优势。然而,基于元球的方法在物理准确性和切割模拟上存在不足。鉴于无网格模型在解决软体形变时通常具有较高的精度，本课题将对元球和无网格模型混合的软组织建模展开研究，尝试提出一种兼具两者优势，能有效构建可交互软组织模型的新方法。具体内容包括：个性化人脑的多尺度元球模型自动生成；基于元球和无网格方法混合的物理模型构建；基于元球和无网格混合模型的切割、撕裂、电凝等基本手术操作的模拟等。在此基础上，研制基于元球和无网格混合模型的虚拟手术引擎，并应用到神经外科的手术模拟中。

With the constant application of virtual reality in medicine, VR surgery simulation and planning have attracted increasing interests from clinic, especially in neurosurgery. The core research content in virtual surgery is the human tissue modeling and surgical simulation. Metaball, as a special implicit surface modeling method, can effectively express smooth and blobby surfaces. Since human cerebral cortex has a large number of sulci and gyri, it has natural advantages in handling collision detection and real-time physical deformation, if these complex geometric structures in brain can be expressed and simplified by metaballs. However, metaballs-based methods are weak in handling the accurate physical deformation and cutting simulation. Considering meshfree methods generally have the high physical accuracy, the aim of this project is to research the interactive soft tissue modeling based on metaballs-meshfree hybrid method and explore its application in neurosurgery simulation. The research topics includes: automatic generation of multi-scale metaballs model for the 3d individualized human brain; physical modeling based on metaballs-meshfree hybrid method; cutting, fracture simulation and electric coagulation based on metaballs-meshfree hybrid model. Finally, the metaballs-meshfree hybrid model based virtual surgery engine will be developed, and applied in the VR neurosurgery training.