Scribd
Upload
83 views1 page

Estimating Diffusion Parameters From Polarized Spherical Gradient Illumination

This document proposes a technique to estimate spatially varying diffusion parameters that model subsurface light transport from observations of translucent materials under polarized spherical gradient illumination. It shows that the diffusion constant and diffuse albedo can be directly obtained from the relationship between the observed diffuse albedo, estimated diffuse normal, and polarized spherical gradients without explicit fitting. This allows dense per-point scattering parameters to be acquired from just four photographs instead of requiring structured lighting patterns.

Uploaded by

ZStepan11
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
83 views1 page

Estimating Diffusion Parameters From Polarized Spherical Gradient Illumination

This document proposes a technique to estimate spatially varying diffusion parameters that model subsurface light transport from observations of translucent materials under polarized spherical gradient illumination. It shows that the diffusion constant and diffuse albedo can be directly obtained from the relationship between the observed diffuse albedo, estimated diffuse normal, and polarized spherical gradients without explicit fitting. This allows dense per-point scattering parameters to be acquired from just four photographs instead of requiring structured lighting patterns.

Uploaded by

ZStepan11
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

Estimating Diffusion Parameters from Polarized Spherical Gradient Illumination

Yufeng Zhu Pieter Peers∗ Paul Debevec Abhijeet Ghosh


USC Institute for Creative Technologies The College of William & Mary∗

(a) Diff. albedo (b) Red normals (c) Green normals (d) Blue normals (e) Translucency (f) Rendering
Figure 1: Estimating spatially varying diffusion parameters from polarized spherical gradient illumination. RGB translucency parameters (e) inferred from diffuse albedo (a) and
RGB diffuse normals (b-d).

Introduction. Accurately modeling and reproducing the appear-


ance of real-world materials is crucial for the production of pho-
toreal imagery of digital scenes and subjects. The appearance of
many common materials is the result of subsurface light transport
that gives rise to the characteristic “soft” appearance and the unique
coloring of such materials. Jensen et al. [2001] introduced the
dipole-diffusion approximation to efficiently model isotropic sub-
surface light transport. The scattering parameters needed to drive
the dipole-diffusion approximation are typically estimated by illu-
minating a homogeneous surface patch with a collimated beam of
light, or in the case of spatially varying translucent materials with
a dense set of structured light patterns. A disadvantage of most
existing techniques is that acquisition time is traded off with spatial
density of the scattering parameters.
Polarized Spherical Gradients. Recently, Ma et al. [2007] pro- (a) Diff. albedo (b) Translucency (c) Rendering
posed a technique to obtain high quality estimates of diffuse and Figure 2: Spatially varying diffusion parameters of material samples estimated us-
specular albedo and photometric normal maps from just eight pho- ing spherical gradient illumination. Top-row: Red wax. Bottom-row: Polished marble.
tograph under four different polarized spherical gradient lighting
conditions. In addition, Ma et al. also proposed a hybrid normal Diffusion from Gradients. Relating Equations (1) and (2) yields a
rendering technique that approximates the soft appearance of sub- mechanism for estimating scattering parameters of dipole diffusion
surface scattering with local shading using measured RGB diffuse from observations of the 1st -order spherical gradients. Specifically,
normals. This suggest a connection between spherical gradient il- a BRDF approximation of Equation 2 relates the observed diffuse
lumination and subsurface scattering. albedo Rd (cross-polarized 0th order spherical statistics) to the norm
of the estimated diffuse normal ~n (cross-polarized 1st order spher-
In this work, we aim to formalize this apparent connection between ical statistics [Ma et al. 2007]) via the diffusion constant D. This
subsurface scattering parameters and observations under spherical leads to the following compact relation between the diffusion con-
gradient illumination of translucent materials based on radiative stant and polarized spherical gradients:
transfer theory [Ishimaru 1978]. In particular, we show that dense Rd
per-surface-point scattering parameters can be directly obtained D≈ . (3)
from observations under spherical gradient illumination (cross- |~n|
polarized to discard specular reflections), without resorting to any
Conclusion. Equation 3 provides a mechanism for directly ob-
explicit fitting of observed scattering profiles.
taining dense spatially-varying diffusion parameters from just four
Background. Light transport in highly scattering translucent mate- observations of translucent materials under polarized spherical gra-
rials can be well approximated by diffusion theory [Ishimaru 1978; dient illumination.
Jensen et al. 2001]. According to radiative transfer theory, diffu-
sion can be accurately approximated by a two-term spherical har- References
monic expansion of radiance:
1 3 I SHIMARU , A. 1978. Wave Propagation and Scattering in Random
L(x, ω ) =
φ (x) + ω ·~
E(x), (1) Media. Academic Press, New York.
4π 4π
where φ (c) is the scalar fluence and ~E(x) is the vector irradi- J ENSEN , H. W., M ARSCHNER , S. R., L EVOY, M., AND H ANRA -
ance. Substituiting Equation 1 in the radiative transfer equation HAN , P. 2001. A practical model for subsurface light transport.
and assuming semi-infinite material, leads to the well-known dif- In Proceedings of ACM SIGGRAPH 2001, 511–518.
fuse BSSRF [Jensen et al. 2001]: M A , W.-C., H AWKINS , T., P EERS , P., C HABERT, C.-F., W EISS ,
(~n · ~E(xo )) M., AND D EBEVEC , P. 2007. Rapid acquisition of specular and
Rd (r) = −D , (2) diffuse normal maps from polarized spherical gradient illumina-
dΦi (xi )
tion. In Rendering Techniques, 183–194.
where r = ||xo − xi ||, and D = 1/3σt0 is the diffusion constant.

You might also like