time the cornea is photographed. 
    
    
slitlamp was used to visually check the adherence of the 
lens to the corneal surface for every patient. 
Comparisons of corneas measured repeatedly and 
comparisons of  Keratocon and  photokeratoscope 
measurements also indicate that the surface of the 
topography is being accurately represented by the 
contact lens. The most difficult aspect to measure is the 
sensitivity of the solution to rapid changes in corneal 
topography. 
The cornea is mapped in a coordinate system defined by 
the object space control. This means that, although an 
X-Y-Z translation of the patients head between 
successive measurements introduces no errors, it is 
important that the patient fixates the same point each 
Figure 4. Photokeratoscope image of a patient with 
conical cornea. Irregular mires can be seen and an 
impression of downward displacement of the cone. This 
cornea was too steep to measure with a keratometer. 
  
  
  
  
0 5mm 
| 1 | 
0.1mm Contour Interval 
Figure 5. Corneal map showing the displacement of the 
cone and the flattening of the cornea. 
The Keratocon was installed at the Tasmanian Lion's Eye 
Diagnostic Centre and tested on a selection of patients with 
corneal abnormalities. Figures 4 and 5 are of a patient with a 
conical left cornea. The contours show the disturbed 
topography, known as a "droopy cone". The cornea is 
flattened above the apex and steeper inferiorly, with temporal 
displacement of the apex. The photokeratoscope provides 
very little information about this patient. The cornea was too 
steep to measure using keratometry. The Keratocon has 
mapped the cornea to its periphery whereas the 
photokeratoscope has imaged less than 50% of the corneal 
surface. 
4. A DIGITAL KERATOCON 
To meet with clinical acceptance, a biomedical measurement 
system must satisfy the following conditions: 
i. it must not discomfort or intimidate the patient, 
ii. it must only require the patient's cooperation for a short 
period of time, 
iii. it must be reliable, 
iv. it must output measurement parameters that suit the 
clinician, 
v. it must present those parameters in a readily 
understandable form, 
vi. it must be simple to use, 
vii. it must be accurate, and 
viii. the time taken to process and present data must suit the 
application. 
The relative importance of each of these conditions will 
depend upon the application but in broad terms, and certainly 
in the case of any method of measuring corneal topography, 
the order in which these have been listed is indicative of their 
importance. 
Accuracy may be the least important of the issues. In the 
case of corneal measurement this is illustrated by the success 
of keratoscopic methods. Speed, except in the case of 
intraoperative procedures, is also not absolutely necessary. 
The turn around time on clinical procedures such as 
angiograms and X-rays indicate the delay that clinicians are 
prepared to tolerate. Ease of use is important, but 
examination of existing ophthalmic instrumentation such as 
keratoscopes and slit lamps suggests that clinicians will 
tolerate quite poor ease of use. The most critical 
determinants of an instrument's success in a normal clinical 
environment appear to be the first four on the list, namely the 
reliable and easily interpreted presentation of appropriate 
information from an instrument that minimises patient 
discomfort. Appropriate information includes estimates of 
how the measured corneal abnormalities will affect vision 
and the information needed in order that consequent vision 
defects can be corrected. 
A digital prototype of the keratocon should therefore meet 
the following criteria: 
i. sufficient metric reliability to avoid the need for on-the- 
job calibrations, 
ii. real time video imagery for patient alignment, 
iii. automatic or semi-automatic target recognition and 
measurement, 
iv. automatic computation of three dimensional data, 
v. high data reliability, 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B5. Vienna 1996 
  
vi. hig 
inf 
vii. ne 
ap| 
The inv 
is a cri 
typical 
reduced 
to signi 
compen 
flow f 
relative 
limit su. 
digital « 
would | 
which n 
15mm 3 
betweer 
framegr 
from a 
techniqi 
photogr 
allow tl 
time pri 
clinical 
therefor 
image. 
In the c 
with 10 
scale 1: 
diamete 
accurac 
than +4 
analogu 
expecte 
  
Fig 
COIT 
Target ; 
aspect o 
has beei 
of autor 
targeted 
Scanned 
The im 
histogra 
is reasc