RIDGE: Re-Noising with Internal Dynamic Guidance for Image Editing

2026-08-03

AI Summary

This work addresses the degradation in editing quality observed in existing inversion-free flow-based image editing methods under high noise levels, which stems from an overly aggressive update due to the constant displacement assumption. To overcome this limitation, we propose a training-free editing approach that dynamically re-noises intermediate latents to maintain proximity to the underlying clean target image and incorporates an internal dynamic guidance mechanism during high-noise stages. Our key innovations include a noise-level-adaptive displacement shrinkage strategy and a soft dynamic mask generated internally by the diffusion model itself, enabling precise spatial control without reliance on external models. Built upon the flow-matching framework, our method is compatible with backbone models such as SD3 Medium and FLUX.1-dev, and achieves a superior balance among source fidelity, target alignment, and perceptual quality on two established benchmarks.

Abstract

Inversion-free flow-based image editing avoids latent inversion, but still requires a target-side state at every editing step. The widely used equal-displacement construction keeps the displacement between the noisy source state and the target-side state unchanged across noise levels. This is inconsistent with noising, under which the displacement between two clean states noised with the same noise level and noise sample should contract as the noise level increases. Thus, it can lead to overly aggressive updates at high noise levels. We introduce RIDGE: Re-Noising with Internal Dynamic Guidance for Image Editing, an inversion-free and training-free method that maintains the edited state as an evolving approximation to the unavailable clean target state. RIDGE re-noises this approximation using the same noise level and noise sample as the clean source state, allowing their noisy displacement to decrease naturally with increasing noise. Since the edited state initially contains limited target semantics, RIDGE further applies internal dynamic guidance during the early high-noise steps. A clean target state prediction guides the provisional edited state through a soft dynamic mask derived internally from the model, focusing guidance on regions that require modification without external segmentation or detection models. Experiments on two benchmarks using two backbones, SD3 Medium and FLUX.1-dev, show that RIDGE offers a favorable aggregate trade-off among source preservation, target alignment, and perceptual quality.

Problem

• flow-based image editing
• inversion-free
• noise displacement
• target-side state
• image editing

Innovation

• inversion-free editing
• re-noising
• dynamic guidance
• flow-based generation
• noise-level consistency