An Empirical Comparison of Monte Carlo Radiosity Algorithms

Philippe Bekaert, Ronald Cools, and Yves D. WIllems.

Computer Graphics Research Group, Department of Computer Science, K. U. Leuven
Celestijnenlaan 200A, B-3001 Heverlee, Belgium

7-th International Conference in Central Europe on Computer Graphics, Visualization and Interactive Digital Media'99, to be held in Plzen, Czech Republic, February 8--12, 1999.


Monte Carlo radiosity algorithms are radiosity algorithms in which the radiosity integral equation or system of linear equations is solved using Monte Carlo random walk techniques. Since explicit form factor computation and storage is completely avoided in Monte Carlo radiosity algorithms, these algorithms are more reliable and require significantly less storage than other radiosity algorithms, making it feasible to render much more complex scenes. This paper presents a comparative study of four main aspects in which proposed Monte Carlo radiosity algorithms differ: whether the discrete or continuous equation is being solved, the random walk state transition simulation technique, sampling order and the sample number generator.

Keywords: Global Illumination, Radiosity, Quasi Monte Carlo, Random Walk

This page is maintained by