Abstract
Purpose: :
To compare fit–error RMS in wavefront fitting using Zernike and Fourier polynomials. To determine if there is an advantage in wavefront representation using a particular polynomial fit.
Methods: :
210 human eye exams (pre and post LASIK, and keratoconic) were measured in a prospective study using a high resolution wavefront sensor (200um pitch). The centroid images were fitted with Zernike (5th order) and Fourier (at least 176 terms). The fit–error RMS, defined as the deviation of the polynomial fitting from the true wavefront elevation calculated for these two surface representations. The mean fit–error RMS for each polynomial fitting were compared for accuracy and precision.
Results: :
Of the 210 exams fitted with Zernike and Fourier polynomials, Zernike fitting resulted in lower fit–error RMS in 95.0% of exams and Fourier fitting in 2.2%. Zernike and Fourier were equivalent (<0.05um difference) in 2.8% of exams. The mean fit–error RMS for Zernike fitting was 0.72um and for Fourier fitting was 1.31um. Zernike fit–error RMS was up to 6.7x lower compared to Fourier in those cases where Zernike fit–error was lower. Fourier fit–error RMS was up to 1.63x lower compared to Zernike in those cases where Fourier fit–error was lower.
Conclusions: :
Zernike fitting more accurately represented the true wavefront elevation in this sample of pre and post surgical and diseased human eye exams.
Keywords: refractive surgery