Progressive photon mapping (PPM) [1] and their derivatives are state-of-the-art
and the industrial choice of global illumination rendering methods. Although they
can generate a converged image faster than other global illumination algorithms do,
because all photons and collection points are generated independently and the ani-
mation is rendered in a frame-by-frame manner, the temporal and spatial coherence
is not explored. As a result even each frame looks converged but the resulted anima-
tion has annoying flickering artifacts. Those flickering artifacts occur because all the
samples are independent so temporal coherency among samples are not well explored
in PPM. As a result even PPM finds an important sample, it will be discarded to
construct another independent sample.
This thesis proposes a hybrid method, Energy Redistribution Progressive Pho-
ton Mapping (ER-PPM), which uses energy redistribution (ER) to explore the tem-
poral and spatial coherency, and uses PPM to get a fast converged result. ER is
used to generate important samples correlatedly to remove the flickering artifacts
and explore important samples. Furthermore, rendering a photorealistic animation
for a complex scene usually takes days to weeks on a single computer with average
performance. Therefore, in order to speed up the rendering process, a parallel an-
imation rendering system is also created. Using Condor, our rendering system is
designed to render an animation in parallel among computers.
Results demonstrate that our algorithm can maintain the robustness of the
PPM method, while significantly reduce the flickering artifacts and improves the
rendering efficiency for a wide range of scenes with complex lighting.

Progressive photon mapping (PPM) [1] and their derivatives are state-of-the-art
and the industrial choice of global illumination rendering methods. Although they
can generate a converged image faster than other global illumination algorithms do,
because all photons and collection points are generated independently and the ani-
mation is rendered in a frame-by-frame manner, the temporal and spatial coherence
is not explored. As a result even each frame looks converged but the resulted anima-
tion has annoying flickering artifacts. Those flickering artifacts occur because all the
samples are independent so temporal coherency among samples are not well explored
in PPM. As a result even PPM finds an important sample, it will be discarded to
construct another independent sample.
This thesis proposes a hybrid method, Energy Redistribution Progressive Pho-
ton Mapping (ER-PPM), which uses energy redistribution (ER) to explore the tem-
poral and spatial coherency, and uses PPM to get a fast converged result. ER is
used to generate important samples correlatedly to remove the flickering artifacts
and explore important samples. Furthermore, rendering a photorealistic animation
for a complex scene usually takes days to weeks on a single computer with average
performance. Therefore, in order to speed up the rendering process, a parallel an-
imation rendering system is also created. Using Condor, our rendering system is
designed to render an animation in parallel among computers.
Results demonstrate that our algorithm can maintain the robustness of the
PPM method, while significantly reduce the flickering artifacts and improves the
rendering efficiency for a wide range of scenes with complex lighting.

[1] T. Hachisuka, S. Ogaki, and H. W. Jensen, “Progressive photon mapping,” in
ACM SIGGRAPH Asia 2008 papers, SIGGRAPH Asia ’08, (New York, NY,
USA), pp. 130:1–130:8, ACM, 2008.
[2] J. Arvo, “Transfer Functions in Global Illumination,” in SIGGRAPH ’93
Course Notes, pp. 1–28, 1993.
[3] L. Szirmay-Kalos, “Monte Carlo methods for global illumination,” in Spring
Conference of Computer Graphics, pp. 1–28, 1999.
[4] M. Pharr and G. Humphreys, Physically Based Rendering, Second Edition:
From Theory To Implementation. San Francisco, CA, USA: Morgan Kaufmann
Publishers Inc., 2nd ed., 2010.
[5] N. Metropolis and S. Ulam, “The Monte Carlo method,” Journal of the Amer-
ican Statistical Association, vol. 44, pp. 335–341, 1949.
[6] J. M. Hammersley and D. C. Handscomb, Monte Carlo Methods. John Wiley
and Sons, Inc., New York, NY, USA, 1965.
[7] E. Veach and L. J. Guibas, “Metropolis light transport,” in SIGGRAPH ’97,
pp. 65–76, 1997.
[8] D. Cline, J. Talbot, and P. Egbert, “Energy redistribution path tracing,” ACM
Trans. Graph. (SIGGRAPH ’05), vol. 24, no. 3, pp. 1186–1195, 2005.
[9] C. Damez, K. Dmitriev, and K. Myszkowski, “State of the art in global illu-
mination for interactive applications and high-quality animations,” Computer
Graphics Forum, vol. 22, no. 1, pp. 55–77, 2003.
[10] Y.-C. Lai, Photorealistic Animation Rendering with Population Monte Carlo
Eenergy Redistribution. PhD thesis, University of WisconsinMadison, 2009.
46[11] S. Fan, S. Chenney, and Y.-C. Lai, “Metropolis photon sampling with optional
user guidance,” in Proc. of the 16th Eurographics Symposium on Rendering,
pp. 127–138, 2005.
[12] T. Hachisuka and H. W. Jensen, “Stochastic progressive photon mapping,” in
ACM SIGGRAPH Asia 2009 papers, SIGGRAPH Asia ’09, (New York, NY,
USA), pp. 141:1–141:8, ACM, 2009.
[13] J. Chen, B. Wang, and J.-H. Yong, “Improved Stochastic Progressive Photon
Mapping with Metropolis Sampling,” (Prague, Czech Republic), pp. 1205–1213,
Eurographics Association, 2011.
[14] E. Veach, Robust Monte Carlo Methods for Light Transport Simulation. PhD
thesis, Stanford University, 1997.
[15] Condor, “http:://www.cs.wisc.edu/condor.”
[16] D. Thain, T. Tannenbaum, and M. Livny, “Distributed computing in practice:
the condor experience.,” Concurrency - Practice and Experience, vol. 17, no. 2-
4, pp. 323–356, 2005.