Nonlinearly Weighted First-order Regression for Denoising Monte Carlo Renderings

1Disney Research 2Walt Disney Animation Studios 3Edinburgh Napier University 4Dartmouth College

In Computer Graphics Forum (Proceedings of EGSR), 2016

Teaser
Top: Visual comparison of our denoising algorithm (NFOR) against previous work on one of the 21 scenes included in our test bench. The input is rendered using 256 spp, each technique uses the same set of features. Bottom: Quantitative comparison showing means and standard deviations of four metrics (MSE and rMSE; lower is better, PSNR and SSIM; higher is better) computed over 64, 256, and 1024 spp renderings of the 21 scenes in the test bench; each blue line corresponds to one scene/spp configuration, all values are expressed relative to the best denosing technique thereof.

Abstract

We address the problem of denoising Monte Carlo renderings by studying existing approaches and proposing a new algorithm that yields state-of-the-art performance on a wide range of scenes. We analyze existing approaches from a theoretical and empirical point of view, relating the strengths and limitations of their corresponding components with an emphasis on production requirements. The observations of our analysis instruct the design of our new filter that offers high-quality results and stable performance. A key observation of our analysis is that using auxiliary buffers (normal, albedo, etc.) to compute the regression weights greatly improves the robustness of zero-order models, but can be detrimental to first-order models. Consequently, our filter performs a first-order regression leveraging a rich set of auxiliary buffers only when fitting the data, and, unlike recent works, considers the pixel color alone when computing the regression weights. We further improve the quality of our output by using a collaborative denoising scheme. Lastly, we introduce a general mean squared error estimator, which can handle the collaborative nature of our filter and its nonlinear weights, to automatically set the bandwidth of our regression kernel.

Downloads

Supplemental Video

Text Reference

Benedikt Bitterli, Fabrice Rousselle, Bochang Moon, José A. Iglesias-Guitián, David Adler, Kenny Mitchell, Wojciech Jarosz, Jan Novák. Nonlinearly Weighted First-order Regression for Denoising Monte Carlo Renderings. Computer Graphics Forum (Proceedings of EGSR), 35(4):107–117, June 2016.

BibTex Reference

@article{bitterli16nonlinearly,
    author = "Bitterli, Benedikt and Rousselle, Fabrice and Moon, Bochang and Iglesias-Guitián, José A. and Adler, David and Mitchell, Kenny and Jarosz, Wojciech and Novák, Jan",
    title = "Nonlinearly Weighted First-order Regression for Denoising Monte Carlo Renderings",
    journal = "Computer Graphics Forum (Proceedings of EGSR)",
    volume = "35",
    number = "4",
    pages = "107–117",
    month = "jun",
    year = "2016",
    doi = "10.1111/cgf.12954"
}

Copyright Disclaimer

© The Author(s). This is the author's version of the work. It is posted here by permission of The Eurographics Association for your personal use. Not for redistribution. The definitive version is available at diglib.eg.org.