2
$\begingroup$

Real human anatomies look wet/ moist when operated. I am looking for some good methods for rendering anatomies in a virtual simulation, and hence I have done a literature survey on 'Rendering wet materials' using keywords like:

  1. Rendering wet materials.
  2. Rendering moist surfaces.
  3. Rendering wet surfaces. etc.

What I observed is, in general, people use ray-tracing methods after modeling surface with a two-layer surface reflection model, for achieving realism for wet surfaces.

What are some good rasterization methods/ approximations for the same keeping in mind real-time performance and unavailability of RTX gfx cards, as well as realism?

$\endgroup$
  • $\begingroup$ You do not necessarily need raytracing to simulate wet surfaces. It is certainly the most realistic approach, but the important factor here is the reflectivity of the material. A water film on skin has a mirror-like reflection behavior, while the microstructure of the skin scatters light into all directions. So pure skin looks much duller than wet skin. In a Phong model you would just modify the specular term of wet skin to get a more reflective surface. In physically-based rendering approaches, you would decrease the microfacet roughness (not sure if this is the right term ;) ) $\endgroup$ – wychmaster Jun 21 at 22:07
  • $\begingroup$ There are also special techniques to simulate the "partial transparency of skin". Just search for "subsurface scattering". You might want to take a look into the rendering section of this wikipedia articel to get started. $\endgroup$ – wychmaster Jun 21 at 22:11
  • $\begingroup$ I am not trying to render skin here but some organs like the brain/liver. SSS approximates material properties but what I need is surface scattering approximation. $\endgroup$ – Argha Chakraborty Jun 22 at 5:11
3
$\begingroup$

The reason why ray-tracing is often used for achieving as-realistic-as-possible wet appearance is their reflective behavior. A thin liquid film on the surface reflects photons like a mirror. If you target real-time rendering, you will have to fake the reflections (or rely on RTX :) There are two ways I can quickly think of: placing reflection probes in the scene using cube-maps or using screen space reflections.

Reflection probes with cube-maps

  • good for static scenes, you can pre-compute many of them
  • allow objects outside of the current viewing frustum to be seen in the mirrors
  • inaccurate due to limited number of probes

There is a great extensive tutorial by Sébastien Lagarde on reflections which continues (what a coincidence) with part B on wet surfaces and there is part C as well. Make sure to check them out!

Screen-space reflections

  • fine also for dynamic scenes
  • only objects in the current viewing frustum can be seen in the mirrors
  • accurate within the frustum

Screen space reflections are explained as part of this awesome series by David Lettier.

Subsurface scattering

Regarding the discussion on subsurface scattering, any organic tissue exposes this property. We just observe it the most on skin or leaves, but the same effect is observed inside the body. Just read section 3.3 of this paper:

One significant casual effect under endoscopic environment is a strong subsurface scattering on the surface of internal organs.

I suggest to use at least some post-processing technique for that.

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.