Geometrical approximations for accurate evaluation of refraction in the human cornea

Corneal analysis is usually done under two assumptions: first it is considered as a single surface that separates air from a medium with an equivalent keratometric index. The analysis is done under exact ray tracing, which leads to a non-uniformly sampled output pattern. In this paper we analyze both of these affirmations. We will first study the importance of the second corneal surface. We also propose several simplifications to the problem of calculating the transmittance of the cornea without resampling the output pattern. The different approximations will be analyzed by computing the differences in the optical path length covered by the light when passing the cornea. We will show that unique evaluation if absolute differences in the optical path is not the best criterion, since it fails in describing the spherical aberration. We will also analyze the absolute optical path function by decomposing it in Zernike polynomials. This criterion will lead us to determine whether approximations proposed are good enough to describe the cornea.