RESUMO
Virtual surgery provides important assistance in preoperative planning of brain surgeries. A virtual surgery system essentially consists of visualization and modeling of the medical data and deformation simulation of human organs (mainly soft tissues). We used the hybrid level set theory, threshold segmentation, and morphological methods to segment the human head MRI volumetric data into five parts, namely the outline of the head, gray matter, white matter, cerebrospinal fluid and ventricles. The iso2mesh toolkit based on Delaunay algorithm was then employed to generate the tetrahedral mesh and mark the five parts. We further studied the soft tissue deformation, and established a mechanic model based on Tensor-Mass system to simulate the point-touch and ball-touch models for soft tissue deformation. This approach can help to improve the positioning accuracy of a virtual brain surgery and represent the process of brain tissue deformation.