Is there currently anyone working on a link between these two fields? I have a hunch that, due to the repetitive nature of sampling, there could be a good bit to gain in the area of ray tracing using Machine Learning techniques, but I'm interested in finding out more and getting involved if there's a group out there doing this.
1
-
1$\begingroup$ I have seen a few people apply machine learning to BRDF's. Here's one such thing: iaeng.org/IJCS/issues_v42/issue_1/IJCS_42_1_04.pdf $\endgroup$– Alan WolfeCommented Jan 5, 2017 at 5:05
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
A group from UC Santa Barbara published the Siggraph 2015 paper A Machine Learning Approach for Filtering Monte Carlo Noise describing a technique which trains a neural network to select filtering parameters for path tracing. The paper details the structure of the neural network, the filter parameters used and how several secondary features are pre-computed to feed the neural network.