Introduction
Human vision module is an Ocean™ 2022 brand-new feature. This module, available as a post-processing filter, can reproduce several properties of the human eyes that would alter the appearance of a scene, depending on the visual abilities of the observers. This module can be used for projects in automotive to evaluate the impact of aging on the perception of interior and exterior light sources such as screens, dashboard lights or streetlights.
Besides, since Ocean™ simulations are visualized on screens with limited color and light reproduction capacities, we use human based tone-mapping functions to adjust the colors of the images when using human vision module. It brings and reinforces a realistic aspect of a given scene while it is displayed on a screen compared to a real-life experience. In addition, the following properties have been (or will be) added to Ocean™:
- Difference between night (grey) and day (color) vision
- Eye sensitivity that changes with age
- Glare effect due to pupil and cornea
- Time adaptation to change of luminosity
Human base tone-mapping
Tone mapping is a technique used in computer graphics to map the high dynamic range (HDR) of an image to a lower dynamic range (LDR) that can be displayed on a screen or printed. This is necessary because most display devices, such as monitors and printers, have a limited dynamic range that cannot display the full range of brightness values captured in an HDR image.
Human-based tone mapping is a technique that aims to mimic the way the human visual system perceives brightness and contrast. The human visual system has the ability to adapt to different lighting conditions and can perceive a wider range of brightness values than most display devices. Therefore, by using a tone mapping technique that is based on the way the human visual system works, the resulting LDR image can look more natural and realistic.
In Ocean™, a first human based tone mapping was chosen because of its simplicity and because it shown a correct conservation of image colorimetry. It is presented in detail in this paper : Adaptive Logarithmic Mapping for Displaying High Contrast Scenes F. Drago, K. Myszkowski, T. Annen, and N. Chiba (In Eurographics 2003), and will be call “Drago” tone mapping in Ocean™ and in the rest of this article.
The basic idea behind the “Drago” method is to apply a non-linear transformation to the pixel values in the image. This transformation compresses the dynamic range of the image while preserving as much detail as possible. It use logarithmic transformation of the luminance value of the input image and control the compression applied to the image thanks to two parameters : b and Ld. Those parameters need to be chosen by the user :
- b : contrast adjustment parameters, between 0 and 1, with a best value of 0.85.
- Ld : maximum luminance that can be displayed by the seen.
The “Drago” tone-mapping node is available as Output shown in figure 1. An example of used is shown in figure 2, where a driving night scene is shown with and without Drago tone mapping. Without tone-mapping, car exterior is saturated, while with tone-mapping, it is more visible.
Figure 1 – Human Based Tone Mapping filter
Figure 2 – Night driving scene with (right) and without Drago (left)
Glare
Glare is a phenomenon that occurs when the brightness of a light source exceeds the level to which the eyes have adapted. It is often experienced as discomfort or visual impairment and can be caused by a variety of sources, including the sun, headlights, reflections off water or snow, and artificial lighting. Glare is a common problem, especially for individuals who drive at night, work in brightly lit environments, or spend time outdoors in bright sunlight. It can cause eye strain, headaches, and reduced visual acuity, making it difficult to see objects clearly and accurately.
There are two types of glare: disability glare and discomfort glare. Disability glare occurs when the brightness of a light source interferes with the ability to see an object clearly. This can be particularly dangerous when driving, as it can make it difficult to see road signs, pedestrians, or other vehicles. Discomfort glare, on the other hand, occurs when the brightness of a light source causes visual discomfort without necessarily interfering with visual acuity.
Individual differences in eye anatomy and physiology can also affect the level of glare experienced. For example, older adults may be more susceptible to glare due to changes in the lens of the eye that occur with age.
The simulation of disability glare in Ocean™ is fully based on this paper : Physically-Based Glare Effects for Digital Images – Greg Spencer and al.
Figure 3 - Point Spread Function principle
The glare simulation is based on the definition of a human eye Point Spread Function (PSF). In general, a point spread function (PSF) describes the response of an imaging system to a point source or point object and is presented in figure 3.
In case of human eye, the PSF, shown in figure 4, is calculated in the section 3 of the paper: Physically-Based Glare Effects for Digital Images – Greg Spencer and al. which is based on J.Vos PSF definition that has attempted to unify the large number of PSF models for the eye.
Figure 4 - Photopic and scotopic PSF (extract from Physically-Based Glare Effects for Digital Images - Greg Spencer and al )
This PSF depends on the parameters of the optical system considered. As the eye adapt to the luminosity (e.g: pupil radius decrea7ses with luminosity, cone cells are active only under well-lit conditions,… ), its parameters might vary and so the PSF. This node defines three kinds of PSF dedicated to three different lighting conditions (see figure 5):
- Photopic vision (10 to 10 8 cd/m²) – Day light
- Mesopic vision (0.01 to 3.0 cd/m²) – Twilight
- Scotopic vision (10 -6 to 10 -3 cd/m²) – Night
Figure 5 - Scotopic, mesopic and photopic regimes definition
Furthermore, the ratio of scattered light on unscattered light increase with age which leads to more glare. The PSF is applied on pixel that satisfy this condition: I(X,Y,Z) > I(Y) . threshold
where is I(X,Y,Z) a 3 channels pixel (CIE XYZ), I(Y) is the average luminance of the input image and threshold is user defined (with default value being 10). This threshold allows to applied the PSF only on light source (otherwise glare effect would be applied on the entire image). A age condition is added to the filter, since glare increase with age.
For photopic conditions, glare is visible as a halo around light source, which is as large as the light source is in the center of the field of view. For scotopic and mesopic conditions, in addition to the halo around light source, a lenticular colored halo (due to dispersion) and random straight lines are also visible, as shown in figure 6.
Figure 6 - Photopic glare halo (left) and scotopic glare halo showing colored halo and straight lines (from : Physically-Based Glare Effects for Digital image)
In Ocean™, the glare filter include three parameters that need to be set by the user (lighting conditions, dispersion, age and threshold (see figure 7).
Figure 7 - Glare filter
An example of glare filter applied to a night (scotopic) scene is shown in figures 8 (without filter) and 9 (with filter). The glare filter applied is a scotopic filter, for a human of 40 years, with dispersion.
Figure 8 - Night scene without glare filter
Figure 9 - Night Scene with glare filter (scotopic for a human of 40 years)
Conclusion
Human vision module in Ocean™ is available as post-processing filter, allowing a direct comparison of simulation with and without human vision effects. Two filters are available : human based tone-mapping and glare filter. More advanced features on Human Vision are in development and will be available soon, such as :
- Difference between night (grey) and day (color) vision
- Eye sensitivity that changes with age
- Time adaptation to change of luminosity
Responses