Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2512.16864 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914208102219776 |
|---|---|
| author | Qu, Tianyuan Ke, Lei Zhan, Xiaohang Tang, Longxiang Liu, Yuqi Peng, Bohao Yu, Bei Yu, Dong Jia, Jiaya |
| author_facet | Qu, Tianyuan Ke, Lei Zhan, Xiaohang Tang, Longxiang Liu, Yuqi Peng, Bohao Yu, Bei Yu, Dong Jia, Jiaya |
| contents | Instruction-based image editing enables natural-language control over visual modifications, yet existing models falter under Instruction-Visual Complexity (IV-Complexity), where intricate instructions meet cluttered or ambiguous scenes. We introduce RePlan (Region-aligned Planning), a plan-then-execute framework that couples a vision-language planner with a diffusion editor. The planner decomposes instructions via step-by-step reasoning and explicitly grounds them to target regions; the editor then applies changes using a training-free attention-region injection mechanism, enabling precise, parallel multi-region edits without iterative inpainting. To strengthen planning, we apply GRPO-based reinforcement learning using 1K instruction-only examples, yielding substantial gains in reasoning fidelity and format reliability. We further present IV-Edit, a benchmark focused on fine-grained grounding and knowledge-intensive edits. Across IV-Complex settings, RePlan consistently outperforms strong baselines trained on far larger datasets, improving regional precision and overall fidelity. Our project page: https://replan-iv-edit.github.io |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_16864 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | RePlan: Reasoning-guided Region Planning for Complex Instruction-based Image Editing Qu, Tianyuan Ke, Lei Zhan, Xiaohang Tang, Longxiang Liu, Yuqi Peng, Bohao Yu, Bei Yu, Dong Jia, Jiaya Computer Vision and Pattern Recognition Instruction-based image editing enables natural-language control over visual modifications, yet existing models falter under Instruction-Visual Complexity (IV-Complexity), where intricate instructions meet cluttered or ambiguous scenes. We introduce RePlan (Region-aligned Planning), a plan-then-execute framework that couples a vision-language planner with a diffusion editor. The planner decomposes instructions via step-by-step reasoning and explicitly grounds them to target regions; the editor then applies changes using a training-free attention-region injection mechanism, enabling precise, parallel multi-region edits without iterative inpainting. To strengthen planning, we apply GRPO-based reinforcement learning using 1K instruction-only examples, yielding substantial gains in reasoning fidelity and format reliability. We further present IV-Edit, a benchmark focused on fine-grained grounding and knowledge-intensive edits. Across IV-Complex settings, RePlan consistently outperforms strong baselines trained on far larger datasets, improving regional precision and overall fidelity. Our project page: https://replan-iv-edit.github.io |
| title | RePlan: Reasoning-guided Region Planning for Complex Instruction-based Image Editing |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.16864 |