Research map / Image generation / Text-to-image diffusion
SmartBrush: Text and Shape Guided Object Inpainting With Diffusion Model
Shaoan Xie, Zhifei Zhang, Zhe Lin, Tobias Hinz, Kun 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 |
|---|---|---|---|
| Train with instance masks and local text descriptions rather than random erased regions and global captions | SmartBrush · OpenImages and MSCOCO, bounding box and object layout masks | Local FID: OpenImages: 12.57 -> 9.71 (box), 10.98 -> 7.82 (layout); MSCOCO: 18.13 -> 13.22 (box), 15.16 -> 9.80 (layout) | Improved |
| Train with instance masks and local text descriptions rather than random erased regions and global captions | SmartBrush · OpenImages and MSCOCO, bounding box and object layout masks | FID: OpenImages: 7.07 -> 6.00 (box), 5.84 -> 4.70 (layout); MSCOCO: 8.50 -> 8.05 (box), 6.54 -> 5.76 (layout) | Improved |
| Train with instance masks and local text descriptions rather than random erased regions and global captions | SmartBrush · OpenImages and MSCOCO, bounding box and object layout masks | CLIP Score: OpenImages: 0.264 -> 0.266 (box), 0.261 -> 0.263 (layout); MSCOCO: 0.246 -> 0.252 (box), 0.243 -> 0.249 (layout) | Mixed |
| Add object-mask prediction and use the predicted mask during sampling (ablation) | SmartBrush · Coarse-mask inpainting | background preservation | Improved |
| Vary mask precision from accurate object masks to coarse masks (ablation) | SmartBrush · Text- and shape-guided inpainting | mask-shape controllability | Improved |
Later papers that compared with this method
2 later accepted papers report SmartBrush as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Improving Text-guided Object Inpainting with Semantic Pre-inpainting (ECCV 2024 · better on 8 of 9 shared results)
- A Task is Worth One Word: Learning with Task Prompts for High-Quality Versatile Image Inpainting (ECCV 2024 · better on 6 of 7 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
- Paint by Inpaint: Learning to Add Image Objects by Removing Them First (CVPR 2025 · similarity 0.90)
- MTADiffusion: Mask Text Alignment Diffusion Model for Object Inpainting (CVPR 2025 · similarity 0.89)
- HD-Painter: High-Resolution and Prompt-Faithful Text-Guided Image Inpainting with Diffusion Models (ICLR 2025 · similarity 0.89)
- Brush2Prompt: Contextual Prompt Generator for Object Inpainting (CVPR 2024 · similarity 0.89)
- TurboFill: Adapting Few-step Text-to-image Model for Fast Image Inpainting (CVPR 2025 · similarity 0.88)
- Towards Enhanced Image Inpainting: Mitigating Unwanted Object Insertion and Preserving Color Consistency (CVPR 2025 · similarity 0.88)
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.
