Beyond Fixed Inference: Quantitative Flow Matching for Adaptive Image Denoising

2026-04-02

AI Summary

Existing image denoising methods suffer from limited performance under unknown or varying noise levels, primarily due to a mismatch between the noise distributions assumed during training and those encountered during inference. To address this, this work proposes a noise-adaptive flow matching framework that leverages local pixel statistics to estimate the input noise level in real time and dynamically adjusts the inference trajectory—including the starting point, number of integration steps, and step size—to align with the actual noise characteristics. This approach is the first to couple quantitative noise estimation with flow matching inference, thereby breaking away from fixed inference paradigms. It demonstrates exceptional robustness and generalization across natural, medical, and microscopic images, significantly improving both denoising accuracy and computational efficiency.

Abstract

Diffusion and flow-based generative models have shown strong potential for image restoration. However, image denoising under unknown and varying noise conditions remains challenging, because the learned vector fields may become inconsistent across different noise levels, leading to degraded restoration quality under mismatch between training and inference. To address this issue, we propose a quantitative flow matching framework for adaptive image denoising. The method first estimates the input noise level from local pixel statistics, and then uses this quantitative estimate to adapt the inference trajectory, including the starting point, the number of integration steps, and the step-size schedule. In this way, the denoising process is better aligned with the actual corruption level of each input, reducing unnecessary computation for lightly corrupted images while providing sufficient refinement for heavily degraded ones. By coupling quantitative noise estimation with noise-adaptive flow inference, the proposed method improves both restoration accuracy and inference efficiency. Extensive experiments on natural, medical, and microscopy images demonstrate its robustness and strong generalization across diverse noise levels and imaging conditions.

Problem

• image denoising
• unknown noise
• noise level mismatch
• adaptive inference
• flow-based generative models

Innovation

• Quantitative Flow Matching
• Adaptive Denoising
• Noise Level Estimation
• Flow-based Generative Models
• Inference Efficiency