# Noisy-As-Clean: Learning Self-supervised Denoising from the Corrupted Image

## TL;DR

This paper introduces a novel self-supervised denoising strategy called Noisy-As-Clean (NAC), which trains networks directly on corrupted images by treating them as clean targets, effectively addressing domain gap issues in image denoising.

## Contribution

The proposed NAC method enables training denoising networks solely on corrupted images, achieving comparable or superior results to supervised methods without requiring clean image pairs.

## Key findings

- NAC achieves competitive denoising performance on synthetic and real noise.
- Networks trained with NAC outperform or match existing supervised and unsupervised methods.
- NAC simplifies training by using corrupted images as targets, reducing reliance on clean datasets.

## Abstract

Supervised deep networks have achieved promisingperformance on image denoising, by learning image priors andnoise statistics on plenty pairs of noisy and clean images. Unsupervised denoising networks are trained with only noisy images. However, for an unseen corrupted image, both supervised andunsupervised networks ignore either its particular image prior, the noise statistics, or both. That is, the networks learned from external images inherently suffer from a domain gap problem: the image priors and noise statistics are very different between the training and test images. This problem becomes more clear when dealing with the signal dependent realistic noise. To circumvent this problem, in this work, we propose a novel "Noisy-As-Clean" (NAC) strategy of training self-supervised denoising networks. Specifically, the corrupted test image is directly taken as the "clean" target, while the inputs are synthetic images consisted of this corrupted image and a second and similar corruption. A simple but useful observation on our NAC is: as long as the noise is weak, it is feasible to learn a self-supervised network only with the corrupted image, approximating the optimal parameters of a supervised network learned with pairs of noisy and clean images. Experiments on synthetic and realistic noise removal demonstrate that, the DnCNN and ResNet networks trained with our self-supervised NAC strategy achieve comparable or better performance than the original ones and previous supervised/unsupervised/self-supervised networks. The code is publicly available at https://github.com/csjunxu/Noisy-As-Clean.