Learning to Recorrupt: Noise Distribution Agnostic Self-Supervised Image Denoising

TL;DR

Learning to Recorrupt (L2R) is a novel self-supervised image denoising method that does not require prior noise distribution knowledge, using a learnable neural network to adapt to various complex noise types.

Contribution

Introduces a noise distribution-agnostic denoising approach with a learnable recorruption process via a min-max objective, outperforming existing methods on diverse noise models.

Findings

Achieves state-of-the-art results on heavy-tailed and correlated noise.

Effectively handles signal-dependent noise like Poisson-Gaussian.

Eliminates the need for prior noise distribution knowledge.

Abstract

Self-supervised image denoising methods have traditionally relied on either architectural constraints or specialized loss functions that require prior knowledge of the noise distribution to avoid the trivial identity mapping. Among these, approaches such as Noisier2Noise or Recorrupted2Recorrupted, create training pairs by adding synthetic noise to the noisy images. While effective, these recorruption-based approaches require precise knowledge of the noise distribution, which is often not available. We present Learning to Recorrupt (L2R), a noise distribution-agnostic denoising technique that eliminates the need for knowledge of the noise distribution. Our method introduces a learnable monotonic neural network that learns the recorruption process through a min-max saddle-point objective. The proposed method achieves state-of-the-art performance across unconventional and heavy-tailed noise distributions, such as log-gamma, Laplace, and spatially correlated noise, as well as signal-dependent noise models such as Poisson-Gaussian noise.