Research map / Image generation / Editing and style transfer
EEdit : Rethinking the Spatial and Temporal Redundancy for Efficient Image Editing
Zexuan Yan, Yue Ma, Chang Zou, Wenteng Chen, Qifeng Chen, Linfeng 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 |
|---|---|---|---|
| add spatial locality caching with inversion step skipping and TIP | FLUX RF-inversion editing · Cache comparison | latency speedup: 2.46x latency, 3.53x FLOPs | Improved |
| add token index preprocessing (TIP) to SLoC (ablation) | SLoC · Prompt editing | inference time: 5.96 -> 5.14 | Improved |
| add inversion step skipping (ISS) to SLoC (ablation) | SLoC · Prompt editing | inference time: 5.96 -> 5.06 | Improved |
| add TIP and ISS together to SLoC (ablation) | SLoC · Prompt editing | inference time: 5.96 -> 4.60 | Improved |
| add TIP and ISS to SLoC (ablation) | SLoC · Dragging | inference time: 7.12 -> 5.66 | Improved |
| add TIP and ISS to SLoC (ablation) | SLoC · Composition | inference time: 7.25 -> 5.66 | Improved |
| add TIP and ISS to SLoC (ablation) | SLoC · Dragging | FG fidelity FID: 20.61 -> 22.07 | Hurt |
| add TIP and ISS to SLoC (ablation) | SLoC · Prompt editing | FG fidelity FID: 39.50 -> 39.21 | Improved |
| skip inversion steps (2-step skip) instead of full inversion (ablation) | SLoC · Prompt editing | inference time: 13.27 -> 10.16 | Improved |
| increase inversion skip interval to 4 (ablation) | SLoC · Prompt editing | inference time: 13.27 -> 8.76 | Improved |
| skip denoising steps instead of inversion steps (ablation) | SLoC · Prompt editing | FG fidelity FID: 3.23 -> 64.10 | Hurt |
| use SLoC token scoring instead of ToCa/DuCa caching | FLUX editing · Cache comparison | FG FID: 84.77 -> 39.50 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Flow-Edit — FlowEdit: Inversion-Free Text-Based Editing Using Pre-Trained Flow Models (ICCV 2025) · better on 2 of 2 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
- Object-Aware Inversion and Reassembly for Image Editing (ICLR 2024 · similarity 0.88)
- Move and Act: Enhanced Object Manipulation and Background Integrity for Image Editing (AAAI 2025 · similarity 0.86)
- Inversion-Free Image Editing with Language-Guided Diffusion Models (CVPR 2024 · similarity 0.86)
- PnP Inversion: Boosting Diffusion-based Editing with 3 Lines of Code (ICLR 2024 · similarity 0.86)
- Balancing Preservation and Modification: A Region and Semantic Aware Metric for Instruction-Based Image Editing (ICML 2025 · similarity 0.86)
- A Video is Worth 256 Bases: Spatial-Temporal Expectation-Maximization Inversion for Zero-Shot Video Editing (CVPR 2024 · similarity 0.86)
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.
