Research And Development Of The Virtual Surgical Planning System For Orthognathic Surgery

In orthognathic surgery, the traditional 2D surgical planning method can't provide precise 3D structure information of patient's skull and hence make it difficult for surgeons to predict surgery effects and make precise surgical plans. To solve the problem, we developed a virtual surgical planning system for orthognathic surgery. This system can generate 3D head model to let surgeons easily simulate their surgery on the virtual patient model and make prediction of the postoperative effects. In this thesis, we introduce this system and discuss the relevant volume visualization techniques and problem of collision detection in virtual surgery.The system is based on volume visualization techniques. Volume visualization is about the processing, manipulation and visualization of volume data. In this paper, we emphasize on the problems of segmentation and reconstruction of volume data. In volume data segmentation, the threshold based and region based methods are separately discussed. The bone tissues in the volume data which is reconstructed from CT slices are successfully segmented by threshold segmentation method. Then a 3D seed and flood algorithm that is more efficient than traditional 3D filling method is presented. In the chapter of volume reconstruction, the reconstruction methods based on slices and voxels are discussed with emphasis on the iso-surface generation method. We found that the iso-surface generated by using Marching Cubes algorithm on segmented binary volume data is not smooth. So we approach a method to solve the problem by implementing dilation operation on segmented volume data.Collision detection is a difficult problem of virtual surgery. We introduce a few methods to solve the problem and discuss the collision detection in the process of soft tissue cutting and movement of bone tissues. Finally, we approach a collision detection method based on volume data model to detect collisions between bone tissues.The system is designed using structural method. We introduce the architecture and functional components of the system. We mainly discuss the system data structure, methods of 3D environment interaction and virtual osteotomy. In the end, we approach some directions for future work.