Background: Optical aberrations in the optical system may downgrade image quality and cannot be fully compensated by spherocylindrical glasses. The subjectively evaluated visual acuity may be significantly reduced. The purpose of this study was to calculate the image forming properties of the eye using a spotlight source or alternatively extended objects.
Methods: A convex and first derivative continuous (C1) surface from the rough height data of the anterior corneal surface (TMS-1, Tomey, Erlangen) or the anterior and posterior corneal surface (Orbscan, Orbtec, USA) was calculated by means of an interpolating subdivision scheme (modified Butterfly algorithm). The characteristics of the residual refractive surfaces were used according to Navarro's eye model. The focal distance was calculated from the exact raytracing calculation (Snellius' law) of the point-spread function by minimising the variance of the point-spread function. The diffraction property of the aperture stop was implemented with a transmission characteristic according to a radially symmetrical Bessel function within the entrance pupil. The algorithm was realised with a C code on the LINUX platform and applied to a normal eye (example 1, TMS-1), an eye with severe keratoconus (example 2, TMS-1) and an eye with corneal scars (example 3, Orbscan).
Results: The focal distance in example 1 (22.5 mm, 22.6 mm, and 22.8 mm) increased with the pupil diameter (2 mm, 3 mm, and 5 mm). The variance of the approximately radially symmetrical point-spread function in the focal plane attained a minimum value with a pupil size of 3 mm (0.164, 0.104, and 0.230). In example 2, the focal distance changed inconclusively (21.1 mm, 21.0 mm, and 21.3 mm) with the pupil size (2 mm, 3 mm, and 5 mm). The variance of the markedly asymmetrical point-spread function in the focal plane was systematically higher compared to the values of example 1 and reached a minimum value with a pupil size of 3 mm (0.255, 0.224, and 0.371). The imaging of the sinus-modulated pattern is anisotropic due to the asymmetry of the point-spread function. In example 3, the focal distance (22.3 mm, 22.3 mm, and 22.5 mm) did not change systematically with the pupil size (2 mm, 3 mm, and 5 mm). The variance of the nearly radially symmetrical point-spread function changed only marginally between pupil sizes of 2 mm and 3 mm (0.231, 0.239, and 0.338).
Conclusions: Raytracing of corneal topography height data based on refined eye models with the option of auto-focussing has the potential to trace the optical resolution of the eye for arbitrary objects. Further studies on contrast sensitivity and the conversion of the real image to a perceived image by the retina and brain are required for complete modeling of subjective visual acuity.