An improved algorithm for sampling distribution based on MLT
HE Huai-qing
CHEN Shuai
LIU Hao-han
JI Yu
Abstract：In order to solve the random noise problem in the image generation process caused by the sparsity of random sampling distribution in Metropolis light tracking ( MLT) algorithm, a\ mutation strategy based on sampling distribution was proposed. The sampling process was divided into two stages to improve the MLT algorithm, and the mutation strategy in the second stage was affected by the sampling distribution matrix generated from the first stage. In addition, a small scale of sampling was performed at the positions of eight-neighbor pixels of current sampling point. While satisfying the detailed balance condition, the average ratio of scalar contribution function of multiple sample points was set as the acceptance probability. The results show that the improved algorithm can generate smaller noise in the image compared with the original method at the same time, and the effect is equivalent to the better improved algorithms. The problem of big noise in the indirect illumination scene can be solved through improving the mutation strategy.
Keywords：Metropolis light trackingsampling distributionglobal illuminationMarkov chain Monte Carlophotorealistic renderingmutation strategyacceptance probabilitypath selection
Publication Date：2017-01-01
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：8( 646-653 )