The modern computer-aided surgery comes up with increasingly innovative solutions that provide relief for surgeons in their work, or his new opportunity for intervention. This includes augmented reality systems or surgical robotics, such as those developed in the context of this SFB. The use of these systems in the operating room is linked to corresponding methods of preparation and planning, thus defining new requirements for planning systems. Only an intuitively accessible planning system can meet these requirements. It must be of the planning system on the one hand, new methods of surgical planning be made available and the other a safe and timely treatment can be. The overall aim should be to speed up the planning through an automation to make them even intraoperatively usable.

The aim of this project is the development of a surgical planning system whicht meets the requirements above. As a key project, it combines the methods of image acquisition and processing with those of intraoperative implementation using Augmented Reality and surgical robotics.

Because the already existing planning tool is to restrictive concering the integration ability a new tool should be developed. Through modular approaches it should become an open surgical planning and control system. Important factor is a structured user interface for an optimal approach with a simultaneous integration of quality assurance. In addition data interfaces should be redefined and made extensible to new intraoperatove tools such as Augmented Reality provided by sub-project K4.

The robot system RobaCKa (robotic assisted surgery system Karlsruhe), which was developed in the two phases before for automatic milling trajectories of the skull was used in the application phase of 2002-2004 for the first time on a patient in a clinical study. This includes several important aspects like safety, risk analysis, planning, and close interdisciplinary cooperation between engineers and surgeons. After establishing a test environment, the tests on dead animals and the recording of lines on the skin of volunteers, the application for a clinical study of ethics committee and regional council is done. As a necessary prerequisite a systematic risk analysis, which includes the design of hardware and software and the operation took place. Essential aspects of a safe surgical robot system are in addition to the intuitive operation slow and predictable movements of the robot, multi-sensor surveillance, a well-defined system architecture and good cooperation between engineers and medical doctors during the operation. The interdisciplinary cooperation does not begin in the operating room, but with the imaging. After the acquistion and processing of CT images the treatment has to be planned. The whole planning process is done by other subprojects. The location of the robot to the patient is essential, because otherwise complex trajectories cannot be executed. In the end the plan is trasferred to the robot control and the intraoperativ execution follows the points:

1. Fixation of the patient and implantation of the Rigid-Bodies
2. Coverage of the robot and positioned in space
3. Registration
4. Recording the trajectory with sterile color
5. Milling of the trajectory.

In this subproject, techniques and systems of Augmented Reality (AR) are developed and clinically evaluated. The goal is the visualization of preoperatively obtained information directly from the surgical field of operation so that these data represent the view of the surgeon as a unit with the patient. The data included here to visualize both patient-specific anatomical structures and pathologies as well as the results of preoperative planning and simulation.

Besides the development of suitable hardware to produce the software technical requirements for the clinical use of prototypes. These include the calibration of optical overlay used in each case, the correct registry data superimposed with the patient and the detection and correction of occlusion in augmented reality. Other aspects include cooperation with intraoperative imaging or robotics, adapting to changed circumstances during the operation or the intraoperative anatomic review of the design in real time.

This sub-project was created in 2001 by combining all activities of the Collaborative Research on Augmented Reality. At that time there were already two approaches to projector based and glasses based Augmented Reality, which had been developed in the sub-projects Q1 ( "target representation") and Q5 ( "Intra-presentation") in the years 1999 to 2001.

Each part in this project developed system will be first tested in a laboratory environment and then clinically evaluated. Initial tests in the operating room will be conducted on a phantom skull and aim the calibration of all system to specific circumstances. Following a positive vote of the ethics committee of Heidelberg University is needed to work with volunteers and later on with selected pations under real clinical conditions.

The 3D sonography has been proved by investigations conducted in recent years as a reliable and logistically available procedure that can effectively improve the range of imaging in diagnosis, therapy planning and follow-up of bone metastases. The acquisition of a volume data block allows, among others, reliable and clinically relevant analysis of extensive processes and the traceability of the objective results.

The other focus of this project is on evaluating and improving the generation of surface models. There are two types of modeling can be distinguished:

Firstly, the creation and evaluation of individual surface models from real data of a patient: The software tool "patient model" allows the transparent control of the process for creating three-dimensional models from image data sets. These were first disclosed the individual processing steps of the process chain and documented transparently for the user with the help of the developed language of documentation "KaPML. Segmentation, 3D visualization, decimation, smoothing, etc. are presented in the form of a flowchart. The parameterisation of the algorithms can be optimized so specifically in terms of precision. Then the tool Kanalyze allows control of the quality of the created models from image data sets. A graphic-visual data comparison is possible in the form of difference models as well as a quantitative comparison.

Second, the creation and evaluation of ideal surface models from records of several patients: the creation of reference models for a specific operation planning was established in another software package, and evaluated using a number of training records. The quality of the reference database is growing continuously with the increase of the number of underlying records. With the established methods of preparation of this reference models and the model comparison, methods for the target / actual comparison of pre-, intra-and postoperative data are provided. Thus, the necessary quality assurance for clinical applications of new methods is possible in other sub-projects.