Sunday, January 20, 2008
Saturday, January 19, 2008
The great challenge of my project is to increase the overall accuracy of the robotic setup. For future validation of the system sub-millimeter accuracy has to be achieved. Let us introduce the major components of the neurosurgical system (click on the side figure for details). Patients are undergoing intensive CT scanning before the surgery. Usually we have 0.5 mm slices, sometimes 2 mm (for research purposes that may be enough). Presently the DICOM format CT scans are transformed into the open source 3D Slicer 3.0 format, therefore we can display them on our computers. The Virtual Fixtures are also defined in Slicer, by putting together the adequate planes. The CT scan is downloaded to the StealthStation platform that has to be registered to the actual skull by touching at least four distinguished points on the skull (in vivo these are screws in the bone). This procedure is the same with every optical tracking device, and the best achievable accuracy is 0.5-0.9 mm. Now we can read the relative position of the robot’s marker to the marker mounted on the patient (or on the operating table).
Another registration procedure is to determine the relation between the virtual frame at the robot’s base (derived from the optical marker mounted on the robot’s end) and the StealthStation’s inner coordinate system. We can calculate this by moving the robot into six different positions, where we can read the end’s position through the StealthLink interface, and calculate the base position by applying the (inverse) forward kinematics based on the joint values provided by the robot encoders. The accuracy of the procedure depends a lot on the range of the joints we move the robot, as we do not have the absolute precise kinematic model of the manipulator. The encoders’ accuracy allows a maximum of 0.15 mm precision.
Finally we have to determine the cutter tool’s tip spatial position compared to both the marker and the tool center point of the NeuroMate. This we achieve by performing pivot calibration, using again the forward kinematics. We approach the same point (the bottom of a cone) in 6 different orientations, therefore we can compute the position of the tooltip.
After all, we have the transformations to go from any frame to any other. As of now, we use the robot’s world frames to determine the motion commands, but later we will relocate it to the reference frame. In the mean time we are continuously looking for better methods to increase the overall accuracy. The system is pretty complex, so there is a lot we can do.
Tuesday, January 8, 2008
The other feature of the club is that people have to present once a month an important or relevant paper in 15 minutes. Balázs gave the first talk on the single most important comparison paper of computational stereo (finding corresponding image features). The title of the article is A Taxonomy and Evaluation of Dense Two-Frame Stereo Correspondence Algorithms. It evaluates 20 different methods and gives detailed analysis to each technique.
For further CIS materials, I also recommend to visit the Lab seminars' page from last semester.
Monday, January 7, 2008
Though it's too late now to submit any work to the next MMVR, it might worth to keep a eye on the scientific results presented at Long Beach in late January at 16th Medicine meets Virtual Reality. Europe will have its 3rd MIRA in the mean time in Rome, or you may prefer IROS in Nice. If I manage to finish my article, I will present my latest results on extreme telementoring at ATA 2008 in April.
You still have some time to apply for the 14th Nordic – Baltic Conference on Biomedical Engineering and Medical Physics in Riga, or for the 22nd Computer Assisted Radiology and Surgery in Barcelona. Afterwards, we can get ready for the next (30th ) EMBC in Vancouver or for the MICCAI in New York.
For further conferences, visit this site.
As I'm also interested in space research, please take a look at a selection of some space-related events. Next highlight is the 9th iSAIRAS in LA, and you can still apply for the 37th COSPAR Scientific Assembly. The 26th AIAA Applied Aerodynamics Conference may also attract people to Hawaii, and so the 59th International Astronautical Congress in Glasgow.
Happy paper-reading, paper-writing!