2-Shots in the Dark: Low-Light Denoising with Minimal Data Acquisition

TL;DR

This paper introduces a practical noise synthesis method for low-light image denoising that requires minimal data, enabling high-quality training with only one noisy image and one dark frame per ISO setting.

Contribution

The authors propose a novel noise synthesis approach using minimal data, combining Poisson modeling and Fourier-domain spectral sampling, outperforming existing methods.

Findings

Achieves state-of-the-art denoising performance on benchmarks.

Requires only one noisy image and one dark frame per ISO.

Does not rely on large datasets or simplified models.

Abstract

Raw images taken in low-light conditions are very noisy due to low photon count and sensor noise. Learning-based denoisers have the potential to reconstruct high-quality images. For training, however, these denoisers require large paired datasets of clean and noisy images, which are difficult to collect. Noise synthesis is an alternative to large-scale data acquisition: given a clean image, we can synthesize a realistic noisy counterpart. In this work, we propose a general and practical noise synthesis method that requires only one single noisy image and one single dark frame per ISO setting. We represent signal-dependent noise with a Poisson distribution and introduce a Fourier-domain spectral sampling algorithm to accurately model signal-independent noise. The latter generates diverse noise realizations that maintain the spatial and statistical properties of real sensor noise. As opposed to competing approaches, our method neither relies on simplified parametric models nor on large sets of clean-noisy image pairs. Our synthesis method is not only accurate and practical, it also leads to state-of-the-art performances on multiple low-light denoising benchmarks.