Corneal surface reconstruction algorithm using Zernike polynomial representation: improvements.

Recently Sicam et al. [J. Opt. Soc. Am. A21, 1300 (2004)] presented a new corneal reconstruction algorithm for estimating corneal sag by Zernike polynomials. An equivalent but simpler derivation of the model equations is presented. The algorithm is tested on a sphere, a conic, and a toric. These tests reveal significant height errors that accrue with distance from the corneal apex. Additional postprocessing steps are introduced to circumvent these errors. A consistent and significant reduction in height errors is observed across the test surfaces. Finally, Sicam used the conic p-value p as a measure of algorithm efficacy. Further investigation shows that the finite Zernike representation affected the reported results. The p-value should therefore be used with caution as an efficacy measure.