Nonlinear Noise2Noise for Efficient Monte Carlo Denoiser Training

TL;DR

This paper introduces a theoretical framework for applying certain nonlinear functions in Noise2Noise training, enabling effective denoising of HDR images with minimal bias, even when training data is noisy.

Contribution

It develops a class of nonlinear functions with minimal bias for Noise2Noise training and demonstrates their effectiveness in denoising HDR images from Monte Carlo rendering.

Findings

Nonlinear functions can be used in Noise2Noise training with minimal bias.

The method effectively denoises HDR images using only noisy data.

Results approach those of traditional methods using clean reference images.

Abstract

The Noise2Noise method allows for training machine learning-based denoisers with pairs of input and target images where both the input and target can be noisy. This removes the need for training with clean target images, which can be difficult to obtain. However, Noise2Noise training has a major limitation: nonlinear functions applied to the noisy targets will skew the results. This bias occurs because the nonlinearity makes the expected value of the noisy targets different from the clean target image. Since nonlinear functions are common in image processing, avoiding them limits the types of preprocessing that can be performed on the noisy targets. Our main insight is that certain nonlinear functions can be applied to the noisy targets without adding significant bias to the results. We develop a theoretical framework for analyzing the effects of these nonlinearities, and describe a class of nonlinear functions with minimal bias. We demonstrate our method on the denoising of high dynamic range (HDR) images produced by Monte Carlo rendering. Noise2Noise training can have trouble with HDR images, where the training process is overwhelmed by outliers and performs poorly. We consider a commonly used method of addressing these training issues: applying a nonlinear tone mapping function to the model output and target images to reduce their dynamic range. This method was previously thought to be incompatible with Noise2Noise training because of the nonlinearities involved. We show that certain combinations of loss functions and tone mapping functions can reduce the effect of outliers while introducing minimal bias. We apply our method to an existing machine learning-based Monte Carlo denoiser, where the original implementation was trained with high-sample count reference images. Our results approach those of the original implementation, but are produced using only noisy training data.