Research map / Image generation / Editing and style transfer
InstantEdit: Text-Guided Few-Step Image Editing with Piecewise Rectified Flow
Yiming Gong, Zhen Zhu, Minjia Zhang
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use PerRFI inversion instead of DDIM inversion (ablation) | InstantEdit · PIE Bench reconstruction | PSNR: 18.83 -> 21.10 | Improved |
| use PerRFI inversion instead of DDIM inversion (ablation) | InstantEdit · PIE Bench reconstruction | structure distance x10^3: 79.42 -> 48.51 | Improved |
| use Inversion Latent Injection instead of Noise Latent Injection (ablation) | InstantEdit · PIE Bench (4-step) | PSNR: 27.02 -> 27.96 | Improved |
| use Inversion Latent Injection instead of Noise Latent Injection (ablation) | InstantEdit · PIE Bench (4-step) | CLIPScore edited: 22.32 -> 22.82 | Improved |
| use Disentangled Prompt Guidance instead of Pseudo-Guidance (ablation) | InstantEdit · PIE Bench (4-step) | PSNR: 24.19 -> 27.96 | Improved |
| use Disentangled Prompt Guidance instead of Pseudo-Guidance (ablation) | InstantEdit · PIE Bench (4-step) | structure distance x10^3: 19.86 -> 17.14 | Improved |
| add attention masking to DPG (ablation) | InstantEdit · PIE Bench (4-step) | PSNR: 24.79 -> 27.96 | Improved |
| add Canny-conditioned ControlNet (ablation) | InstantEdit · PIE Bench (4-step) | PSNR: 26.68 -> 27.96 | Improved |
| add Canny-conditioned ControlNet (ablation) | InstantEdit · PIE Bench (4-step) | CLIPScore whole: 26.31 -> 26.28 | No clear effect |
| increase generation from 8 NFE to 32 NFE | InstantEdit · PIE Bench | PSNR: 27.96 -> 29.69 | Improved |
| increase generation from 8 NFE to 32 NFE | InstantEdit · PIE Bench | CLIPScore edited: 22.82 -> 22.69 | No clear effect |
| raise ControlNet scale from 0.4 to 0.8 (ablation) | InstantEdit · PIE Bench (4-step) | PSNR vs CLIPScore edited: 27.96 -> 30.17; 22.82 -> 21.52 | Mixed |
Later papers that compared with this method
2 later accepted papers report InstantEdit as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- RewardFlow: Generate Images by Optimizing What You Reward (CVPR 2026 · better on 6 of 7 shared results)
- BiFM: Bidirectional Flow Matching for Few-Step Image Editing and Generation (CVPR 2026 · better on 2 of 4 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
- SwiftEdit: Lightning Fast Text-Guided Image Editing via One-Step Diffusion (CVPR 2025 · similarity 0.92)
- PnP Inversion: Boosting Diffusion-based Editing with 3 Lines of Code (ICLR 2024 · similarity 0.88)
- EditInfinity: Image Editing with Binary-Quantized Generative Models (NeurIPS 2025 · similarity 0.88)
- SplitFlow: Flow Decomposition for Inversion-Free Text-to-Image Editing (NeurIPS 2025 · similarity 0.88)
- FeedEdit: Text-Based Image Editing with Dynamic Feedback Regulation (CVPR 2025 · similarity 0.88)
- Inversion-Free Image Editing with Language-Guided Diffusion Models (CVPR 2024 · similarity 0.87)
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.
