AI Paper Atlas

Research map / Image generation / Editing and style transfer

DirectEdit: Step-Level Accurate Inversion for Flow-Based Image Editing

Desong Yang, Mang Ye

ICML 2026 · Image generation › Editing and style transfer · approach: Instruction-based diffusion editing · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Direct Alignment instead of stepwise correction (ablation)DirectEdit (FLUX.1-dev) · PIE-BenchStructure Distance: 29.22 -> 17.94Improved
add Direct Alignment instead of stepwise correction (ablation)DirectEdit (FLUX.1-dev) · PIE-Benchbackground PSNR: 31.12 -> 32.63Improved
add Direct Alignment (residual alignment of forward path)Euler inversion on FLUX.1-dev · PIE-Bench reconstructionstep-level MSE (avg): 0.2857 -> 0.0006Improved
add Direct Alignmentinversion NFE cost · PIE-Bench reconstructionNFE: 60 -> 60No clear effect
add attention feature injection (ablation)DirectEdit (FLUX.1-dev) · PIE-Benchbackground PSNR: 31.93 -> 32.63Improved
add attention feature injection (ablation)DirectEdit (FLUX.1-dev) · PIE-BenchCLIP similarity (edited): 22.69 -> 22.45Hurt
add mask-guided noise blending (ablation)DirectEdit (FLUX.1-dev) · PIE-Benchbackground PSNR: 24.70 -> 32.63Improved
add mask-guided noise blending (ablation)DirectEdit (FLUX.1-dev) · PIE-BenchCLIP similarity (edited): 22.74 -> 22.45Hurt
use multi-branch mask instead of rectangular bounding box (ablation)DirectEdit (FLUX.1-dev) · PIE-Benchbackground PSNR: 27.86 -> 32.63Improved
add Direct Alignment over vanilla Euler inversion (ablation)DirectEdit (FLUX.1-dev) · PIE-Benchbackground PSNR: 16.81 -> 32.63Improved
increase attention injection steps t_inj (ablation)DirectEdit · PIE-Bench qualitativeedit vs source similarity trade-offMixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 4 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy