A Study On Registration And Preoperative Planning Issues Of Image Guided Surgery System

Image Guided Surgery (IGS), in which computers are used to provide image guidance to physicians for invasive procedures, is still a relatively new field. Most developments have come over the last 10 years. The term of IGS has also been used more broadly to encompass the use of advanced technology in surgery, including surgical planning, surgical robotics, and virtual reality, including surgical simulation. Systems based on bony landmarks for applications such as the brain, spine, and ENT are now commercially available, and some institutions use them as the standard for care. However, systems for abdominal interventions and for other procedures with substantial organ motion and deformation have not yet been developed, and constitute an area of current research.In order to develop the IGS system that could fit for the Chinese technical condition, this thesis does the research which is supported by the National 863 Project as well as the Beijing Municipal Science and Technology Commission foundation. The thesis gives the detailed introduction on the key technologies about the IGS navigation system, including the image acquiring and modeling, preoperative plan and simulation, registration and tracking. In the aspect of image acquiring and modeling, we use volume rendering and surface rendering algorithms to build 3D spine and brain models separately from CT and MRI images. In the aspect of preoperative plan and simulation, we focus on the spinal surgery of pedicle screw fixation surgery, give the detailed analysis on the clinical requirement and functional planning, use 3D segmentation and collision detection techniques to implement the surgical simulation function. In the aspect of registration and tracking, we give the coordinate definition to the components of IGS system. According to the workflow from the clinical navigation surgery, we summarize the registration principle and algorithm, and analyze the capability, precision, as well as influence factors about the registration algorithm. We use Polaris Spectra? which is the renowned real-time 3D optical measurement systems from NID Corporation to combine with the graphic station, and implement the navigation process from detecting the real position of probe to rendering the visual probe model in the graphic station.This thesis is based on the platform of Visual C++, using VTK to construct and render 3D models, using ITK to process and segment images, using MFC to design the interfaces and program the functions, using serial port technique to communicate between the graphic station and Polaris, using the multithreading technique to solve the problem that read data from through serial port, while rendering the probe in the 3D scene.