“Adaptive importance sampling for multi-ray gathering” by Neulander

  • ©Ivan Neulander




    Adaptive importance sampling for multi-ray gathering



    We present an adaptive noise reduction technique for integrating incident radiance at a fixed position. We use a Russian-roulette-based importance sampler to reshape the directional probability density of future rays in a batch, based on an affinity map that incorporates ratings of evaluated rays, provided by the rendering engine. Our method is unbiased, has low overhead, requires no precomputation, and works in concert with other importance sampling schemes.


    1. Pharr, M., and Humphreys, G. 2010. Physically Based Rendering, Second Edition: From Theory To Implementation, 2nd ed. Morgan Kaufmann Publishers Inc., San Francisco, CA, USA.
    2. Veach, E. 1998. Robust monte carlo methods for light transport simulation. PhD thesis, Stanford, CA, USA. AAI9837162.

ACM Digital Library Publication:

Overview Page: