Research map / Image generation / Text-to-image diffusion
The Fabrication of Reality and Fantasy: Scene Generation with LLM-Assisted Prompt Interpretation
Yi Yao, Chan-Feng Hsu, Jhe-Hao Lin, Hongxia Xie, Terence Lin, Yi-Ning Huang, Hong-Han Shuai, Wen-Huang Cheng
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 SAA module on top of guidance and suppression constraints (ablation) | RFNet · RFBench | GPT4Score: 0.572 -> 0.719 | Improved |
| add guidance constraint alone (ablation) | Stable Diffusion with LLM layout · RFBench | GPT4Score: 0.295 -> 0.554 | Improved |
| add suppression constraint alone (ablation) | Stable Diffusion with LLM layout · RFBench | GPT4Score: 0.295 -> 0.407 | Improved |
| add suppression constraint to guidance constraint (ablation) | RFNet · RFBench | GPT4Score: 0.554 -> 0.572 | Improved |
| set guidance strength beta to an intermediate value (15) rather than extreme (0.1 or 30) (ablation) | RFNet · RFBench | qualitative layout and image quality | Improved |
| use RFNet instead of plain Stable Diffusion | Stable Diffusion 2.1 · RFBench (Avg) | GPT4Score: 0.541 -> 0.719 | Improved |
| use RFNet instead of Imagen on DrawBench creativity prompts | RFNet · DrawBench subset | GPT4Score: 0.425 -> 0.125 on blue pizza; 0.194 -> 0.713 on shark in desert | Mixed |
| use GPT4Score instead of CLIPScore as automatic metric | evaluation · RFBench human correlation | Kendall tau: -0.4880 -> 0.6351 on Scientific and Empirical Reasoning | Improved |
Closest papers by meaning
- LLM Blueprint: Enabling Text-to-Image Generation with Complex and Detailed Prompts (ICLR 2024 · similarity 0.89)
- DreamDistribution: Learning Prompt Distribution for Diverse In-distribution Generation (ICLR 2025 · similarity 0.89)
- Mastering Text-to-Image Diffusion: Recaptioning, Planning, and Generating with Multimodal LLMs (ICML 2024 · similarity 0.88)
- Prompt-Free Diffusion: Taking "Text" out of Text-to-Image Diffusion Models (CVPR 2024 · similarity 0.88)
- PromptCoT: Align Prompt Distribution via Adapted Chain-of-Thought (CVPR 2024 · similarity 0.87)
- Enhancing Creative Generation on Stable Diffusion-based Models (CVPR 2025 · 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.
