Research map / Image generation / Text-to-image diffusion
FreLay: Frequency-aware Energy Function for Training-free Layout-to-Image Generation
Bonan Li, Yinhan Hu, Songhua Liu, Zeyu Xiao, Xinchao Wang
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 Anchor Attention Energy Function (ablation) | vanilla attention energy function · COCO2014 | AP: 14.72 -> 19.92 | Improved |
| add Anchor Attention Energy Function (ablation) | vanilla attention energy function · COCO2014 | FID: 27.04 -> 26.55 | Improved |
| add Anchor Attention Energy Function (ablation) | vanilla attention energy function · COCO2014 | CLIP score: 0.328 -> 0.331 | Improved |
| add Timestep Adaptivity Extension (ablation) | Anchor Attention Energy Function · COCO2014 | AP: 19.92 -> 21.91 | Improved |
| add Timestep Adaptivity Extension (ablation) | Anchor Attention Energy Function · COCO2014 | FID: 26.55 -> 26.39 | Improved |
| add Timestep Adaptivity Extension (ablation) | Anchor Attention Energy Function · COCO2014 | CLIP score: 0.331 -> 0.339 | Improved |
| replace WinWinLay with FreLay | WinWinLay · COCO2014 | AP: 19.74 -> 21.91 | Improved |
| replace WinWinLay with FreLay | WinWinLay · Flickr30K | AP: 17.28 -> 18.67 | Improved |
| replace WinWinLay with FreLay | WinWinLay · user study | controllability preference: 22.5 -> 32.9 | Improved |
| replace WinWinLay with FreLay | WinWinLay · user study | quality preference: 25.5 -> 26.6 | Improved |
| increase low-frequency window size n (ablation) | FreLay · qualitative | spatial alignment vs detail | Mixed |
Closest papers by meaning
- Control and Realism: Best of Both Worlds in Layout-to-Image without Training (ICML 2025 · similarity 0.86)
- Energy-Guided Optimization for Personalized Image Editing with Pretrained Text-to-Image Diffusion Models (AAAI 2025 · similarity 0.85)
- SpotActor: Training-Free Layout-Controlled Consistent Image Generation (AAAI 2025 · similarity 0.85)
- NoiseCollage: A Layout-Aware Text-to-Image Diffusion Model Based on Noise Cropping and Merging (CVPR 2024 · similarity 0.84)
- Smoothed Energy Guidance: Guiding Diffusion Models with Reduced Energy Curvature of Attention (NeurIPS 2024 · similarity 0.84)
- SEGA: Spectral-Energy Guided Attention for Resolution Extrapolation in Diffusion Transformers (NeurIPS 2026 · similarity 0.84)
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.
