Computer Graphics and Visualization • Publications

Computer Graphics Forum

Geometric Sample Reweighting for Monte Carlo Integration

Jerry Guo and Elmar Eisemann

Wineglass with dispersive dielectric materials with 32 samples per pixel. As can be seen, our reweighted schemes improve upon uniform and stratified samples and produces much smoother result in terms of color noise. Significant improvements can be observed in low discrepency sequences. All results are plotted in log-log scale.

Numerical integration is fundamental in multiple Monte Carlo rendering problems. We present a sample reweighting scheme, including underlying theory, and analysis of numerical performance for the integration of an unknown one-dimensional function. Our method is simple to implement and builds upon the insight to link the weights to a function reconstruction process during integration. We provide proof that our solution is unbiased in one-dimensional cases and consistent in multi-dimensional cases. We illustrate its effectiveness in several use cases.

More Information

Publisher Paper preprint Slides Supplemental

Citation

Jerry Guo and Elmar Eisemann, Geometric Sample Reweighting for Monte Carlo Integration, Computer Graphics Forum, 40, pp. 109–119, 2021.

BibTex

@article{bib:guo:2021,
    author       = { Guo, Jerry and Eisemann, Elmar },    
    title        = { Geometric Sample Reweighting for Monte Carlo Integration },
    journal      = { Computer Graphics Forum },
    volume       = { 40 },
    year         = { 2021 },
    pages        = { 109--119 },
    doi          = { 10.1111/cgf.14405 },
    dblp         = { journals/cgf/GuoE21 },
    url          = { https://publications.graphics.tudelft.nl/papers/84 },
}