Research On The Key Technologies Of Computer Assisted Surgery (CAS) System

Computer Assisted Surgery (CAS) System, which is one of the hottest fields in modern medicine domain, makes sense of locating the targets, making the surgery plan, gaining the success of surgery. It brings great progress to clinic diagnostics and treatments. Based on the advanced surgery mode of CAS, e.g. medical image guiding, navigation and positioning, the problem of targets positioning in traditional surgery is fixed. Besides to improve the surgical equality, the development of CAS provides the possibilities for Micro Invasive Surgery, Tele-surgery and Robot surgery. Thus, the researches of key technologies of CAS system make sense.In this paper, several main works have been accomplished around CAS system:1. Developing, prospect and features of CAS system are introduced.2. The key technologies of CAS system and the executive process of CAS are analyzed.3. According to the Digital Imaging and Communication of Medicine (DICOM) 3.0 standard, the data structures of DICOM 3.0 standard medical image file and DICOM dataset are investigated, then, the loading module of this kind of files is developed.4. The features of the fiducial markers in medical image are investigated. The basic theory of mathematics morphologic is analyzed. Then the fiducial markers in medical image are automatic localized by mathematics morphologic algorithm, the complex, low precision bring by hand working are solved. Next, we propose a new automatic algorithm for 3D detection and localization of the fiducial marker without manual intervention, which provides high 3D localization accuracy and speed, and is independent of the geometry of the marker.5. The latest navigation and position systems are introduced; the optical tracking system is analyzed in detail. Then, several tracking products and the application of tracking system are introduced.6. Data of reconstruction model before surgery and data acquired by tracking system during surgery are registrated. We present two main registration algorithms: a four-point algorithm for finding the least-squares solution based on SVD(singular value decomposition) of matrix and a ICP(iterative closest point) algorithm based on quaternions.