this on a more 
ing accuracy in 
d, with some 
aphic darkroom 
he potential to 
superimposition 
surgery, which 
| be undertaken 
he final digital 
nt solutions can 
the School of 
wvailable which 
yas immobilised 
raphs, by means 
adhesive targets 
> head was kept 
lane horizontal. 
attached to the 
d with its edges 
ie radiograph as 
wall and the X- 
ncipal axis was 
> principal point 
the cephalostat. 
fixed distance 
plane. When in 
vards by a fixed 
ay tube, allowed 
ich that the front 
spot of the X-ray 
cipal axis of the 
ilar to the mid- 
] with the centre 
an exposure Was 
)CS200 camera, 
y format, was 
through a SCSI 
.dobe Photoshop. 
then applied to 
ess, contrast and 
the form of a 
0 dpi using the 
NN? 
  
  
Figure 1 - DCS200 camera mount fixed to X-ray tube. 
HP ScanJet Plus scanner. It was soon apparent, however, that 
the resulting digital image was poor, lacking both contrast and 
brightness, so that the bone structure could not be determined, 
and subsequent improvement of its radiometric characteristics 
in Adobe Photoshop proved very difficult. This problem arose 
because this scanner was designed for use with prints only and, 
even after placing a white sheet of paper behind the radiograph 
at the time of scanning, the result was still unsatisfactory. 
Consequently, it was decided to use the Kodak DCS200 camera 
instead of the scanner to convert the radiograph into a digital 
image. The radiograph was placed on a light table and the 
camera was aligned with its principal axis perpendicular to the 
light table and passing through the centre of the image of the 
ear-rods. The photographic distance was chosen such that the 
radiograph just filled the frame. This method produced a much 
better image. The image was downloaded into Adobe 
Photoshop, changed to monochrome and enhanced as necessary 
to highlight the bone structure. 
The two images imported into Adobe Photoshop, were brought 
to a common scale and then the X-ray superimposed on the 
photograph. To adjust the image scale, the distance between 
two serrations on the steel ruler was determined from the 
measured pixel coordinates taken on the conventional profile 
image. The same distance was then measured on the 
407 
  
Figure 2 - Combined radiograph and digital image of the head. 
radiograph and a scale factor determined which was applied to 
the radiograph so as to bring both images to a common scale. 
The next step involved highlighting the radiograph as a floating 
selection so that it could be moved around over the photograph 
until the position of the centre of the ear-rod matched on both 
images. Some adjustments to the transparency of the overlain 
radiograph were also needed to ensure that the two images 
were both clearly visible. Following this, a rotation of the 
radiograph about the ear-rod was applied until the serrations on 
the ruler matched. Finally, the combined image was brought to 
a known scale (such as 2 pixels = 1 mm) so that any 
measurements made on it would be meaningful. 
This image file may be passed to the oral surgeon who can 
manipulate the file in Adobe Photoshop. The ‘select and 
move’ function enables any portion of the image, such as the 
lower jaw, to be selected and moved around and the effects of 
the movement on the profile can be seen. Once an acceptable 
position is found, the amount of movement from the existing 
position may be measured on the image. 
A high quality hardcopy output can be obtained from a file 
produced at any stage of the work through the use of a dye 
sublimation printer. An example of the final output is shown 
in Figure 2. 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B5. Vienna 1996