We improve the efficiency of quasi-Monte Carlo integro-approximation by using weighted averages of samples instead of the samples themselves. The proposed deterministic algorithm is constructed such that it converges to the solution of the given integro-approximation problem. The improvements and wide applicability of the consistent method are demonstrated by visual evidence in the setting of light transport simulation for photorealistic image synthesis, where the weighted averages correspond to locally smoothed contributions of path space samples.
CITATION STYLE
Keller, A., Dahm, K., & Binder, N. (2016). Path space filtering. In Springer Proceedings in Mathematics and Statistics (Vol. 163, pp. 423–436). Springer New York LLC. https://doi.org/10.1007/978-3-319-33507-0_21
Mendeley helps you to discover research relevant for your work.