Weighted importance sampling techniques for Monte Carlo radiosity


Philippe Bekaert

Katholieke Universiteit Leuven

            

Mateu Sbert

University of Girona

            

Yves Willems

Katholieke Universiteit Leuven



Contact: Computer Graphics Research Group

Proceedings of the Eurographics Workshop on Rendering 2000, p. 35-46
Brno, Czech Republic (26-28 June)





Abstract

This paper presents weighted importance sampling techniques for Monte Carlo form factor computation and for stochastic Jacobi radiosity system solution. Weighted importance sampling is a generalisation of importance sampling. The basic idea is to compute a-posteriori a correction factor to the importance sampling estimates, based on sample weights accumulated during sampling. With proper weights, the correction factor will compensate for statistical fluctuations and lead to a lower mean square error. Although weighted importance sampling is a simple extension to importance sampling, our experiments indicate that it can lead to a substantial reduction of the error at a very low additional computation and storage cost.


Keywords: Radiosity, Monte Carlo method, Weighted Importance Sampling, Variance Reduction, Stochastic Jacobi iterative method, Form Factor Computation


Downloads


Links