Purchase this article with an account.
V.D. P. Sicam, J.E. Coppens, T.J. T. P. van den Berg, G.L. van der Heijde; Performance of a Corneal Topographer that uses Pseudo Random Encoded Stimulus . Invest. Ophthalmol. Vis. Sci. 2004;45(13):2881.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: We show the performance of a corneal topographer that uses pseudo random encoded stimulus pattern. Critical numerical and experimental tests were done to describe the advantage of using this topographer instead of the usual Placido ring system. Methods: The cornea topographer we use differs from traditional topographers that use Placido rings as stimulus pattern. A color–coded stimulus pattern is used instead thus making it possible to uniquely determine one–to–one correspondence between points on the recorded Purkinje image and points on the source pattern using a pseudo random array. The surface reconstruction procedure involves tracing rays from image points to stimulus points and fitting of surface normals to calculate the correct Zernike coefficients representing the corneal surface. Numerical simulations were made which include different surfaces to evaluate the accuracy of the surface reconstruction algorithm in determining Zernike coefficients. Included in these simulations is a test to determine whether the topographer is not affected by skew ray errors resulting from axial symmetry dependence when using Placido rings as targets. Measurements on real eyes were also done and from this measurement noise was calculated. Results: Numerical tests show that the surface reconstruction method used gives Zernike coefficients with an accuracy of better than 0.3 µm. Numerical tests also show that the topographer can recover corneal shapes without being affected by skew ray errors. For real eyes the repeatability in measuring the Zernike coefficients describing corneal anterior surface is better than 2.0 µm. This repeatability can still be improved with better eye–tracking methods to minimize variations in the measurements due to eye movements. The signal to noise ratio obtained from real eye data reveals that measurement variability is comparable to Zernike coefficients of considerable high order. Conclusions: The surface reconstruction method used has the numerical potential of determining surface height values and Zernike coefficients (for corneal shape) with sub micron accuracy. We have further shown that using a colored pattern source eliminates the problem of meridional ambiguity making it a very good alternative compared to the use of traditional placido ring topographers. The numerical error of the algorithm is small compared to measurement variability when measuring real eyes. Thus, further improvement on the accuracy of the surface reconstruction procedure is not necessary.
This PDF is available to Subscribers Only