Real-time realistic skin translucency

Area of Science:

• Computer Graphics
• Computational Imaging
• Physics Simulation

Background:

• Realistic skin rendering is crucial for immersive visual experiences in computer graphics.
• Existing diffusion theory models offer realistic results but can be computationally expensive.
• Screen-space diffusion approximations provide real-time performance but struggle with light transmittance.

Purpose of the Study:

• To develop an efficient, real-time algorithm for photorealistic skin rendering.
• To address the limitations of screen-space diffusion in simulating light transmittance.
• To create a globally efficient solution for both reflectance and transmittance in multilayered skin models.

Main Methods:

• Utilized texture-space diffusion with Gaussian approximation and programmable graphics hardware.
• Implemented a novel transmittance algorithm within the screen-space diffusion framework.
• Derived transmittance calculations from physical equations accessed via textures.

Main Results:

• Achieved real-time, photorealistic skin renderings with accurate light reflectance and transmittance simulation.
• The screen-space approach with the new transmittance algorithm overcomes limitations of previous methods.
• The method requires minimal additional processing power and memory bandwidth.

Conclusions:

• The developed algorithm provides an efficient global solution for realistic skin rendering in real-time applications.
• This technique enables high-fidelity visual effects in performance-critical scenarios like video games.
• The model effectively reproduces visual qualities comparable to offline and real-time rendering techniques such as photon mapping.