Item


Go with the Winners Strategy in Path Tracing

This paper proposes a new random walk strategy that minimizes the variance of the estimate using statistical estimations of local and global features of the scene. Based on the local and global properties, the algorithm decides at each point whether a Russian-roulette like random termination is worth performing, or on the contrary, we should split the path into several child paths. In this sense the algorithm is similar to the go-with-the-winners strategy invented in general Monte Carlo context. However, instead of establishing thresholds to make decisions, we compute the number of child paths on a continuous level and show that Russian roulette can be interpreted as a kind of splitting using fractional number of children. The new method is built into a path tracing algorithm, and a minimum cost heuristic is proposed for choosing the number of re°ected rays. Comparing it with the classical path tracing approach we concluded that the new method reduced the variance signi¯cantly

© Journal of WSCG, 2005, vol. 13, núm. 1-3, p.49-56

University of West Bohemia, Czech Republic

Author: Szirmay-Kalos, László
Date: 2005
Abstract: This paper proposes a new random walk strategy that minimizes the variance of the estimate using statistical estimations of local and global features of the scene. Based on the local and global properties, the algorithm decides at each point whether a Russian-roulette like random termination is worth performing, or on the contrary, we should split the path into several child paths. In this sense the algorithm is similar to the go-with-the-winners strategy invented in general Monte Carlo context. However, instead of establishing thresholds to make decisions, we compute the number of child paths on a continuous level and show that Russian roulette can be interpreted as a kind of splitting using fractional number of children. The new method is built into a path tracing algorithm, and a minimum cost heuristic is proposed for choosing the number of re°ected rays. Comparing it with the classical path tracing approach we concluded that the new method reduced the variance signi¯cantly
Format: application/pdf
ISSN: 1213-6972 (versió paper)
1213-6964 (versió electrònica)
Document access: http://hdl.handle.net/10256/9453
Language: eng
Publisher: University of West Bohemia, Czech Republic
Collection: Reproducció digital del document publicat a: http://wscg.zcu.cz/wscg2005/Papers_2005/Journal/!WSCG2005_Journal_Final.pdf
Articles publicats (D-IMA)
Is part of: © Journal of WSCG, 2005, vol. 13, núm. 1-3, p.49-56
Rights: Tots els drets reservats
Subject: Montecarlo, Mètode de
Monte Carlo method
Estimació, Teoria de l’
Estimation theory
Rutes aleatòries (Matemàtica)
Random walks
Title: Go with the Winners Strategy in Path Tracing
Type: info:eu-repo/semantics/article
Repository: DUGiDocs

Subjects

Authors