Simulating subsurface scattering in skin by near-infrared light Master’s thesis in Computer science and engineering

Abstract

The reflectance properties of human skin in the visible spectrum and near-infrared spectrum are significant. In general, skin appears more uniform. In this thesis we have chosen to explore the difference between subsurface scattering in near-infrared and visible light and also how to implement subsurface scattering in a real-time setting. We measure a diffusion profile of Caucasian skin in near-infrared light at 940 nm on the inside of the forearm. This diffusion profile is used with existing real-time skin rendering algorithms to implement subsurface scattering. Our diffusion profile of 940 nm did not show a major difference to that of red light (700 nm), however, it is not conclusive as measurements of red light were not our own. Further, when implementing subsurface scattering we used previous research and do not introduce any novel implementation. Finally, we conclude that a large part of the increased uniformity can be attributed to the fact that the skin is only being illuminated by a single wavelength (940 nm). Consequently, small details, normally highlighted by absorption in the visible spectrum become far less noticeable. However, nearinfrared light also penetrates deeper into the skin, highlighting other details, e.g., blood vessels in certain areas of the face.