Abstract:
A number of different methods have been proposed for representing and calculating the optical errors of the eye for different pupil sizes. While Zernike polynomials are a convenient, compact notation, they suffer from errors when used for highly aberrated eyes.
Several methods for calculating and representing the wavefront error from aberrometer slope data have been developed. We compared these methods using data acquired with a high dynamic–range aberrometer. By analyzing data from challenging subjects (i.e. post–lasik) we can compare the accuracy of the various algorithms. In this case the measured area includes both modified and unmodified zones. Analysis of pre– and post–lasik data at different pupil sizes allows a quantitative comparison of the various mathematical methods.
For normal eyes, even with significant aberrations, the various methods agree. However, for abnormal eyes (i.e. kerataconus or post–lasik), the results may depend critically on analysis. Specifically, we have found that the Zernike modal reconstructors require a very large fit order, which may be subject to numerical instability. In addition, if the RMS fit error is too large, the Zernike coefficient based resizing and recentering algorithms (proposed in ANSI Z80–28.2004) fail completely. In these cases the zonal reconstruction method was found to still provide an accurate wavefront representation.
It is important to understand the assumptions behind the analysis methods before using the results. Specifically this applies to the use of pupil recentered or resized data that has large residual wavefront fit errors and Zernike amplitudes with large uncertainties.
Keywords: refractive surgery: optical quality • optical properties • refraction