Research map / Image generation / Editing and style transfer
SR-Edit: Region-Aware Image Editing via Self-Refinement
Andong Wang, Zehua Chen, Yuxuan Jiang, Jun Zhu
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 two self-refinement iterations instead of one (ablation) | SR-Edit on Qwen-Image-Edit 2511 · ImgEdit-Bench | SSIM / Judge: SSIM 0.60 -> 0.67; Judge 3.60 -> 3.94 | Improved |
| estimate regions with few-step inference instead of one-step reconstruction (ablation) | SR-Edit · ImgEdit-Bench | SSIM: 0.64 -> 0.67 | Improved |
| use linearly increasing guidance instead of constant (ablation) | SR-Edit · ImgEdit-Bench | SSIM: 0.62 -> 0.67 | Improved |
| use linearly increasing guidance instead of linearly decreasing (ablation) | SR-Edit · ImgEdit-Bench | SSIM: 0.61 -> 0.67 | Improved |
| add SR-Edit to Qwen-Image-Edit 2511 | Qwen-Image-Edit 2511 · ImgEdit-Bench | LPIPS: 0.46 -> 0.31 | Improved |
| add SR-Edit to AnyEdit | AnyEdit · ImgEdit-Bench | SSIM: 0.70 -> 0.85 | Improved |
| add SR-Edit to Step1X-Edit v1p2 | Step1X-Edit v1p2 · ImgEdit-Bench | LPIPS: 0.39 -> 0.27 | Improved |
| add SR-Edit to InstructPix2Pix | InstructPix2Pix · ImgEdit-Bench | CLIP (x100): 26.68 -> 25.07 | Hurt |
| add SR-Edit to Qwen-Image-Edit 2511 | Qwen-Image-Edit 2511 · PIE-Bench | non-edit SSIM: 0.87 -> 0.94 | Improved |
| SR-Edit vs SpotEdit on Qwen-Image-Edit 2511 | SR-Edit · ImgEdit-Bench | SSIM: 0.67 vs 0.70 (SpotEdit) | Hurt |
| add SR-Edit instead of SpotEdit for runtime | SR-Edit · Inference time (Qwen-Image-Edit 2511) | time (s): 53.69 (vanilla) -> 75.31 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- InstructPix2Pix — InstructPix2Pix: Learning To Follow Image Editing Instructions (CVPR 2023) · better on 3 of 3 shared results
- AnyEdit — AnyEdit: Mastering Unified High-Quality Image Editing for Any Idea (CVPR 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
- Improving Editability in Image Generation with Layer-wise Memory (CVPR 2025 · similarity 0.88)
- ImgEdit: A Unified Image Editing Dataset and Benchmark (NeurIPS 2025 · similarity 0.88)
- Balancing Preservation and Modification: A Region and Semantic Aware Metric for Instruction-Based Image Editing (ICML 2025 · similarity 0.87)
- EditAR: Unified Conditional Generation with Autoregressive Models (CVPR 2025 · similarity 0.87)
- EditBridge: Towards Faithful and Efficient Ultra-High-Resolution Image Editing (NeurIPS 2026 · similarity 0.87)
- UltraEdit: Instruction-based Fine-Grained Image Editing at Scale (NeurIPS 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.
