Path Space Regularization for Holistic and Robust Light Transport

Anton S. Kaplanyan and Carsten Dachsbacher

Computer Graphics Forum (Proceedings of Eurographics 2013)

The Kitchen scene illuminated by a point light from outside through glass windows. MCPPM introduces bias uniformly, while MLT does not: here bias becomes noticeable as blurred edges of the floor caustic (red inset). VCM, one of the recent non-Metropolis techniques, generates noisy results due to the inefficiency of stochastic path sampling (compared to exploring the small subset of light paths through the window contributing to the lighting in the image). The two rightmost figures are rendered using original MLT mutations with the recent manifold exploration mutation. The blue inset shows the reflected caustic for which both MCPPM and regularized MLT introduce unavoidable bias; non-regularized MLT cannot handle the caustic reflected off the mirror wall.

Abstract

We propose a simple yet powerful regularization framework for robust light transport simulation. It builds on top of existing unbiased methods and resorts to a consistent estimation using regularization only for paths which cannot be sampled in an unbiased way. To introduce as little bias as possible, we selectively regularize individual interactions along paths, and also derive the regularization consistency conditions. Our approach is compatible with the majority of unbiased methods, e.g. (bidirectional) path tracing and Metropolis light transport (MLT), and only a simple modification is required to adapt existing renderers. We compare to recent unbiased and consistent methods and show examples of scenes with difficult light paths, where regularization is required to account for all illumination features. When coupled with MLT we are able to sample all phenomena, like recent consistent methods, while achieving superior convergence.

Paper (with thumbnails)

Supplementary

Slides

Bibtex

@article{Kaplanyan2013PSR,
   author  = {Kaplanyan, Anton S. and Dachsbacher, Carsten},
   title   = {Path Space Regularization for Holistic and Robust Light Transport},
   journal = {Computer Graphics Forum (Proc. of Eurographics 2013)},
   volume  = {32},
   number  = {2},
   pages   = {63--72},
   year    = {2013}
 }