Poisson2Gaussian: Noise Gaussianization to Enhance Image Denoising

2026-06-22

AI Summary

Real-world imaging systems often exhibit Poisson-mixed noise characterized by signal dependence, heteroscedasticity, and statistical asymmetry, which existing methods struggle to model accurately. This work proposes Poisson2Gaussian (P2G), a novel approach that achieves high-fidelity Gaussianization of complex real-world noise through probability density matching—going beyond low-order moment alignment—to transform it into independent and identically distributed Gaussian noise. Built upon this transformation, the method establishes an unbiased unsupervised denoising framework that requires neither paired clean images nor explicit noise parameter estimation, enabling seamless integration with any off-the-shelf denoiser. Extensive experiments demonstrate that P2G consistently enhances performance across multiple datasets, achieving up to a 0.75 dB PSNR gain under severely non-Gaussian noise conditions and providing universal performance improvement for diverse denoising architectures.

Abstract

The quantum nature of light determines the inherent Poisson stochasticity of photon detection, which is ubiquitous in photography, microscopy, and astronomy. However, our controlled numerical studies reveal that the signal-dependency, heteroscedasticity, and statistical asymmetry of Poisson-mixed noise make it challenging for existing denoisers to learn. In contrast, i.i.d. Gaussian noise, with its statistical independence and symmetric distribution, is easier to model for networks. To address this gap, we propose Poisson2Gaussian (P2G), a noise Gaussianization method that explicitly converts complex real-world noise to i.i.d. Gaussian noise via probability density matching beyond low-order moments. We also design an unbiased denoising framework that synergizes P2G with downstream denoisers, ensuring convergence to the underlying signal without requiring paired clean data or explicit noise parameters. Extensive experiments demonstrate that P2G consistently achieves state-of-the-art performance across diverse datasets. In challenging scenarios where noise strongly deviates from Gaussian statistics, our method improves the PSNR by up to 0.75 dB. Notably, P2G is architecture-agnostic and can provide universal improvements for various denoisers. The source code will be publicly available.

Problem

• Poisson noise
• image denoising
• heteroscedasticity
• signal-dependent noise
• noise modeling

Innovation

• Poisson2Gaussian
• noise Gaussianization
• heteroscedastic noise
• probability density matching
• blind image denoising