Research map / Image generation / Text-to-image diffusion
BoxDiff: Text-to-Image Synthesis with Training-Free Box-Constrained Diffusion
Jinheng Xie, Yuexiang Li, Yawen Huang, Haozhe Liu, Wentian Zhang, Yefeng Zheng, Mike Zheng Shou
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 outer-box constraint to inner-box constraint (ablation) | BoxDiff · box-conditioned synthesis | YOLO AP: 9.8 -> 20.2 | Improved |
| add corner constraint (ablation) | BoxDiff (IB+OB) · box-conditioned synthesis | YOLO AP: 20.2 -> 22.3 | Improved |
| use topk sampling instead of all elements (ablation) | BoxDiff · box-conditioned synthesis | T2I-Sim: AP 24.8 -> 22.3; T2I-Sim 0.3489 -> 0.3513 | Mixed |
| random sampling instead of topk (ablation) | BoxDiff · box-conditioned synthesis | YOLO AP: 22.3 -> 21.4 | Hurt |
| remove sample() from corner constraint (ablation) | BoxDiff · box-conditioned synthesis | YOLO AP: 22.3 -> 7.7 | Hurt |
| add BoxDiff constraints to Stable Diffusion | Stable Diffusion · box-conditioned synthesis | YOLO AP: 2.8 -> 22.3 | Improved |
| plug BoxDiff into GLIGEN | GLIGEN · box-conditioned synthesis | YOLO AP: 29.7 -> 40.2 | Improved |
| give BoxDiff uncommon object combinations or locations | BoxDiff · unusual prompts | qualitative | Hurt |
Closest papers by meaning
- Fast constrained sampling in pre-trained diffusion models (NeurIPS 2025 · similarity 0.91)
- Any-Size-Diffusion: Toward Efficient Text-Driven Synthesis for Any-Size HD Images (AAAI 2024 · similarity 0.90)
- RealCompo: Balancing Realism and Compositionality Improves Text-to-Image Diffusion Models (NeurIPS 2024 · similarity 0.90)
- Prompt-Free Diffusion: Taking "Text" out of Text-to-Image Diffusion Models (CVPR 2024 · similarity 0.90)
- CtrLoRA: An Extensible and Efficient Framework for Controllable Image Generation (ICLR 2025 · similarity 0.90)
- GrounDiT: Grounding Diffusion Transformers via Noisy Patch Transplantation (NeurIPS 2024 · similarity 0.89)
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.
