Table 1 Summary of the methods discussed in this study that utilize ray tracing for vision simulation

Mostafawy et al. [9]

The authors used distributed ray tracing with 3D scenes. The eye was modeled using spherical refractive elements and a flat retina, and optionally extended by corneal refractive zones. Focusing was achieved by modifying the crystalline lens parameters

No aspherical elements were used, the flat retina caused incorrect peripheral blurring, and no spectral tracing was performed

Fink and Micol [10]

This work used Zernike polynomials to represent the refractive eye elements, distributed ray tracing with 2D images as input, and an inverse ray-tracing procedure with an emmetropic eye model to unwarp the retinal images

The intersection tests were very costly, the input was limited to 2D images, and chromatic aberration was ignored

Wu et al. [11]

The authors used aspherical refractive eye elements to improve the peripheral simulation accuracy and bidirectional path tracing with 3D scenes to guarantee that no rays are blocked by the pupil

Chromatic aberration was ignored and the flat retina shape caused incorrect peripheral blurring

Wang and Xiao [12]

This algorithm utilized spectacle lens prescriptions for a fast simulation of myopia, hyperopia, and astigmatism, and a novel simplified geometrical formulation for a stripped-down eye model to efficiently calculate the refracted ray directions

The supported aberration types were very limited and chromatic aberration was ignored

Wei et al. [13]

This method used distributed ray tracing with 3D scenes and a triangle-based human eye model to significantly improve the computational performance of ray-intersection tests

The intersection tests were overall still costly compared to the closed-form analytical approaches

Dias et al. [14]

The authors examined the effects of multiple retina shapes using distributed ray tracing with 3D scenes and the Navarro schematic eye model

Chromatic effects were ignored and the resulting retinal images were not unwarped

Cholewiak et al. [15]

The authors simulated chromatic aberration by modeling the human eye using a lens with a wavelength-dependent focal length

The minimalistic eye model could not simulate other types of aberrations

Lian et al. [16]

The authors described an open-source software package that used distributed ray tracing with 3D scenes and aspherical eye models and simulated diffraction, spectral effects, and cone excitations

The resulting retinal images were not unwarped

Vu et al. [17]

This work used forward ray tracing with 2D images and aspherical eye elements, and a novel GPU-based rasterization approach that substantially reduced the noise levels resulting from ray tracing

Chromatic aberration was ignored and the simulation was limited to 2D images