Purchase this article with an account.
A. de Castro, IV, S. Barbero, S. Marcos; A Reconstruction Technique to Estimate the Gradient-Index Distribution of the Crystalline Lens Using Ray Aberration Data in vivo. Invest. Ophthalmol. Vis. Sci. 2007;48(13):3818.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Human crystalline lens is known to have a gradient index (GRIN) distribution. Several reconstruction techniques have been proposed to estimate GRIN profiles using as input data the slopes of the rays deflected by the lens in a ray tracing set-up. However ray deflections can only be measured in vitro. The present study explores the possibility of estimating the GRIN profile in vivo using transverse ray aberrations.
We have implemented reconstruction algorithms using the optimization toolbox of an optical design program (Zemax). The GRIN profile is modelled with a 2-parameter function and the biometry data of the lens is assumed to be known. We tested two types of input data in our algorithms: cosines of refracted rays by the lens (only available from in vitro measurements), and transverse ray aberrations (available both in vitro and in vivo using retinal spots). The average difference between the model and estimated GRIN profile (RMS) was used as a metric. In order to perform realistic simulations we evaluated the experimental errors in a set-up that includes a 594-nm He-Ne, a laser scanning system (x-y galvanometer), a water cell chamber to place the lens under test, and two CCD cameras: one for viewing lateral ray deflections and the other conjugated to the focal plane to measure retinal transverse ray aberration (spot deviations). Calibrations were done on artificial lenses with known geometry. Image processing routines have been developed for ray detection.
1) Experimental errors in entrance ray coordinates were 11.5±0.8 µm, image processing errors in the estimation of cosines of the refracted rays estimated from ray deflections was 5.7±1.4*10-4, and typical measurement variability in the transverse aberrations was 0.08 mrad. 2) Using the estimated experimental errors of the entrance ray in the simulations, we obtained an RMS reconstruction error of 9.71±7.42*10-6 when using the cosines of the refracted rays and an RMS of 2.85±2.14*10-5 when using the transverse ray aberrations.
Optimization procedures using ray angular deflections or ray transverse aberrations have been proved to recover the GRIN of the lens within the same order of error. These results indicate that if geometrical properties are known, it should be possible to measure the GRIN lens in vivo using laser ray tracing.
This PDF is available to Subscribers Only