Preciseedit improves image edits by balancing strength and stability
PreciseEdit: Strong Yet Stable Semantic Transport for One-Step Image Editing
Computer Vision and Pattern Recognition
Summary
Editing pictures by changing their meaning in one step is tricky because strong changes can make the image look distorted or accidentally change parts you don't want to. The authors found that strong changes can be done safely if applied carefully to only reliable parts of the image. They designed a new method called PreciseEdit that better matches the desired change while keeping the rest of the image intact, without needing extra training or reversing previous edits. Tests show PreciseEdit works better than other methods at a variety of image editing tasks.
What this means in practice
- •For web developers: Create interactive web-based tools that let users edit images with strong but stable semantic changes in a single step.
- •For graphic design teams: Improve software that applies large conceptual edits to images without altering backgrounds or causing distortions.
Authors
Jingyi Wang, Xiao Lu
Abstract
Recent one-step text-to-image (T2I) models enable efficient image synthesis, opening up new possibilities for interactive Web-based image editing. However, existing one-step editing methods often construct a global transport field through simple vector operations between conditional predictions. While such strong transport can effectively drive the image toward the target semantics, it may be released along unreliable directions or over non-edited regions, leading to structural distortions within the editing region and background drift elsewhere. More importantly, stabilizing the editing process through global smoothing or uniformly weakening the transport field often makes large semantic changes overly conservative. In this work, we revisit one-step image editing from the perspective of the tension between semantic strength and transport stability, and identify a key insight: strong semantic transport is not inherently unstable. What matters is whether it is applied along reliable directions and at reliable spatial locations. Motivated by this insight, we introduce PreciseEdit, a model-agnostic, training-free and inversion-free one-step editing framework that strengthens target-semantic alignment while preserving background fidelity and structural consistency. We further propose a dual-noise-scale role decomposition strategy: a higher-noise scale produces the strong target-directed transport used for actual editing, whereas a lower-noise scale jointly serves as a probe for directional agreement and spatial structure. Extensive experiments on PIE-Bench demonstrate that PreciseEdit achieves state-of-the-art performance and effectiveness on various editing tasks.