Surface-based rigid registration for image-guided surgery

Abstract

Image-guided surgery nowadays is used widely, either directly or indirectly, in the operating room. There is a need for improvement in the registration system to eliminate invasive implants and improve patient comfort. This dissertation proposes a new automatic surface-based rigid registration system for image-guided surgery based on the employment of neural network techniques. specialized in bone structures. The system takes advantage of the time gap between the preoperative scan and interoperative scan in building a patient-specific statistical surface model for preoperative data. The surface model is then used for intraoperative registration. Since no point correspondence is required in this model, the intraoperative registration process is significantly faster.