Noise2Info: Noisy Image to Information of Noise for Self-Supervised Image Denoising

Abstract

Unsupervised image denoising has been proposed to alleviate the widespread noise problem without requiring clean images. Existing works mainly follow the self-supervised way, which tries to reconstruct each pixel x of noisy images without the knowledge of x. More recently, some pioneer works further emphasize the importance of x and propose to weigh the information extracted from x and other pixels when recovering x. However, such a method is highly sensitive to the standard deviation σ <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</inf> of noise injected to clean images, where σ <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</inf> is inaccessible without knowing clean images. Thus, it is unrealistic to assume that σ <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</inf> is known for pursuing high model performance.To alleviate this issue, we propose Noise2Info to extract the critical information, the standard deviation σ <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</inf> of injected noise, only based on the noisy images. Specifically, we first theoretically provide an upper bound on σ <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</inf> , while the bound requires clean images. Then, we propose a novel method to estimate the bound of σ <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</inf> by only using noisy images. Besides, we prove that the difference between our estimation with the true deviation goes smaller as the model training. Empirical studies show that Noise2Info is effective and robust on benchmark data sets and closely estimates the standard deviation of noise during model training.

Topics & Concepts