Research map / Image generation / Editing and style transfer
Multi-turn Consistent Image Editing
Zijun Zhou, Yingying Deng, Xiangyu He, Weiming Dong, Fan Tang
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 multi-turn composite LQR target (vs single-LQR) (ablation) | Ours · PIE-Bench fourth turn | FID: 9.886 -> 5.553 | Improved |
| add multi-turn composite LQR target (vs single-LQR) (ablation) | Ours · PIE-Bench fourth turn | CLIP-T: 26.484 -> 26.831 | Mixed |
| add attention injection (vs w/o attn) (ablation) | Ours · PIE-Bench fourth turn | FID: 6.678 -> 5.553 | Improved |
| add attention injection (vs w/o attn) (ablation) | Ours · PIE-Bench fourth turn | CLIP-I: 0.889 -> 0.894 | Improved |
| use default rather than high attention weights (ablation) | Ours · PIE-Bench fourth turn | FID: 6.316 -> 5.553 | Improved |
| use default rather than high attention weights (ablation) | Ours · PIE-Bench fourth turn | CLIP-T: 26.878 -> 26.831 | No clear effect |
| use Ours instead of RF-Inversion | RF-Inversion · PIE-Bench fourth turn | CLIP-T: 24.094 -> 26.831 | Improved |
| use Ours instead of FlowEdit | FlowEdit · PIE-Bench fourth turn | FID: 14.547 -> 5.553 | Improved |
| use Ours instead of RF-Solver | RF-Solver · PIE-Bench fourth turn | FID: 11.581 -> 5.553 | Improved |
| use Ours (8 steps) instead of 15 steps | Ours · PIE-Bench fourth turn | FID: 5.553 -> 5.396 | Improved |
| use Ours instead of FireFlow-v | FireFlow-v · PIE-Bench fourth turn | CLIP-T: 28.281 -> 26.831 | Hurt |
| apply multi-turn LQR in first editing round | Ours · PIE-Bench first turn | FID: 2.554 vs 0.998 (FlowEdit) | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- StableFlow — Stable Flow: Vital Layers for Training-Free Image Editing (CVPR 2025) · better on 3 of 3 shared results
- FireFlow — FireFlow: Fast Inversion of Rectified Flow for Image Semantic Editing (ICML 2025) · better on 3 of 3 shared results
- FlowEdit — FlowEdit: Inversion-Free Text-Based Editing Using Pre-Trained Flow Models (ICCV 2025) · better on 3 of 3 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 0.87)
- Stable Flow: Vital Layers for Training-Free Image Editing (CVPR 2025 · similarity 0.86)
- Move and Act: Enhanced Object Manipulation and Background Integrity for Image Editing (AAAI 2025 · similarity 0.86)
- Tuning-Free Inversion-Enhanced Control for Consistent Image Editing (AAAI 2024 · similarity 0.86)
- EdiVal-Agent: An Object-Centric Framework for Automated, Fine-Grained Evaluation of Multi-Turn Editing (ICLR 2026 · similarity 0.86)
- Unveil Inversion and Invariance in Flow Transformer for Versatile Image Editing (CVPR 2025 · 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.
