Fig. 1

Part a shows an example of the extreme windowing used for the marker definition. The endpoints can be seen much clearer in the strong windowing. Part b shows X-ray images (window 4500, level 4500) of the ExacTrac system with the predicted marker position in orange and the possible detected positions in blue. The orange circle surrounding the predicted position has a radius of 3 mm. In this image the FM detection was only successful for imager2 (right) with a surrounding intensity (purple circle) of over 5000. The dark borders on the edges come from the partly closed jaws of the imagers. The second marker visible within the image was not defined within the ExacTrac since it is strongly bent. Thus there is no predicted position for this marker