Input-Dependent Uncorrelated Weighting for Monte Carlo Denoising

Jonghee Back¹, Binh-Son Hua², Toshiya Hachisuka³, Bochang Moon¹

GIST¹, Trinity College Dublin², University of Waterloo³

Overview

This code is the official implementation of SIGGRAPH Asia 2023 paper, Input-Dependent Uncorrelated Weighting for Monte Carlo Denoising. For more detailed information, please refer to our project page or other materials as below:

Usage

Running code

First, CMake, CUDA, OpenEXR and zlib are required to run the code. Please clone repositories for OpenEXR and zlib into src/ext folder and rename two folders to openexr and zlib, respectively. After all dependencies are set, please use CMake as follows (or running CMake GUI):

mkdir build
cd build
cmake ..

After building the generated project, you can run the denoiser with some provided example data (refer to below) in the following manner:

./DenoiserTester test_scenes/input-crn test_scenes/reference bathroom crn 128

Please check the print message for more detailed usage:

./DenoiserTester -h

We have tested the code on machines with the following environments:

Windows 10, Visual Studio 2019, CUDA 11.3
Ubuntu 20.04, Clang 10.0.0, CUDA 11.6

Example data

The example data consists of pairs of unbiased independent and correlated pixel estimates (common random number (CRN) and L2 reconstruction for gradient-domain rendering), rendered by gradient-domain rendering framework on top of Mitsuba renderer. The example data can be available here:

License

All source codes are released under a BSD License.

Citation

@inproceedings{Back23,
author = {Back, Jonghee and Hua, Binh-Son and Hachisuka, Toshiya and Moon, Bochang},
title = {Input-Dependent Uncorrelated Weighting for Monte Carlo Denoising},
year = {2023},
isbn = {9798400703157},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3610548.3618177},
doi = {10.1145/3610548.3618177},
booktitle = {SIGGRAPH Asia 2023 Conference Papers},
articleno = {9},
numpages = {10},
keywords = {Monte Carlo denoising, unbiased denoising, input-dependent weighting, uncorrelated weighting},
location = {<conf-loc>, <city>Sydney</city>, <state>NSW</state>, <country>Australia</country>, </conf-loc>},
series = {SA '23}
}

Contact

If there are any questions, issues or comments, please feel free to send an e-mail to jongheeback@gm.gist.ac.kr.

CGLab-GIST/input-dependent-uncorrelated-weighting