Nonlinearly Weighted First-order Regression for Denoising Monte Carlo Renderings

• Published in: Computer graphics forum Vol. 35; no. 4; pp. 107 - 117
• Main Authors: Bitterli, Benedikt, Rousselle, Fabrice, Moon, Bochang, Iglesias-Guitián, José A., Adler, David, Mitchell, Kenny, Jarosz, Wojciech, Novák, Jan
• Format: Journal Article
• Language: English
• Published: Oxford Blackwell Publishing Ltd 01.07.2016
• Subjects: Algorithms; Buffers; Categories and Subject Descriptors (according to ACM CCS); Collaboration; Computer graphics; Computer simulation; I.3.7 [Computer Graphics]: Three-Dimensional Graphics and Realism-Raytracing; I.4.3 [Computer Graphics]: Enhancement-Filtering; Monte Carlo methods; Monte Carlo simulation; Noise reduction; Regression; Regression analysis; Rendering; Studies

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 limit...