Adaptive noise imitation for image denoising

TL;DR

This paper introduces ADANI, an adaptive noise imitation algorithm that synthesizes realistic noisy images from unpaired data, enabling effective supervised denoising without pre-defined noise statistics or paired datasets.

Contribution

The paper proposes a novel adaptive noise imitation method that generates realistic noisy data from unpaired images, facilitating supervised denoising in practical scenarios.

Findings

ADANI produces noise similar to real noisy images both visually and statistically.

Denoising CNN trained with ADANI-generated data achieves competitive performance.

The method works without pre-defined noise models or paired datasets.

Abstract

The effectiveness of existing denoising algorithms typically relies on accurate pre-defined noise statistics or plenty of paired data, which limits their practicality. In this work, we focus on denoising in the more common case where noise statistics and paired data are unavailable. Considering that denoising CNNs require supervision, we develop a new \textbf{adaptive noise imitation (ADANI)} algorithm that can synthesize noisy data from naturally noisy images. To produce realistic noise, a noise generator takes unpaired noisy/clean images as input, where the noisy image is a guide for noise generation. By imposing explicit constraints on the type, level and gradient of noise, the output noise of ADANI will be similar to the guided noise, while keeping the original clean background of the image. Coupling the noisy data output from ADANI with the corresponding ground-truth, a denoising CNN is then trained in a fully-supervised manner. Experiments show that the noisy data produced by ADANI are visually and statistically similar to real ones so that the denoising CNN in our method is competitive to other networks trained with external paired data.