Unifying noise models for improved image denoising with diffusion methods
The Noise Plane: A Diffusion Bridge Across Exponent and Dispersion
Computer Vision and Pattern Recognition
Summary
Images taken by sensors often have different types of noise depending on the sensor and light intensity, not just simple blurry noise. The authors propose a new way to handle many kinds of real-world noise under one framework by using a flexible noise model called Tweedie noise. They create a system that can adapt to various noise types without retraining for each one, making image cleanup more reliable across different scenarios. Their method works well on synthetic images and real data like microscopy and satellite images.
What this means in practice
- •For microscopy image analysts: Improve denoising of fluorescence microscopy images by handling their complex signal-dependent noise with a single adaptable model.
- •For satellite image processors: Enhance synthetic aperture radar (SAR) image cleaning across different sensors by using a unified noise handling approach.
Authors
Xuran Hu, Yuru Jia, Tengxi Wang, Miloš Daković, Ljubiša Stanković, Wufan Zhao
Abstract
Diffusion-based denoising commonly uses Gaussian perturbations, whereas sensor noise follows diverse signal-dependent statistical laws. In this paper, we propose $p$-Bridge, a unified measurement-indexed diffusion framework for Tweedie noise. The noise exponent $p$ describes the variance--intensity relation, while accumulated measurement $ν$ controls the noise level. Together, they define a noise plane that includes Gaussian, Poisson, and Gamma models. The stochastic bridge connects measurement distributions through independent increments, giving denoising a physical interpretation as progressive measurement accumulation. Both the starting and stopping indices can be specified, allowing restoration from an observed measurement level to a prescribed higher level. Measurement superposition yields a shared reverse-mean update, paired with a deviance-based objective for learning the required conditional mean. A latent compound Poisson--Gamma construction enables sampling at intermediate exponents. A single exponent-conditioned network adapts to different Tweedie noise models without separate training for each exponent. Experiments show that exponent conditioning improves cross-exponent robustness over fixed-exponent models. Using estimated exponents, the shared network achieves robustness comparable to an oracle-selected set of five specialist models. Experiments on synthetic images, FMD fluorescence microscopy, and multi-sensor SAR images further demonstrate its effectiveness across different signal-dependent noise regimes.