Research map / Image generation / Text-to-image diffusion
Make-a-Scene: Scene-Based Text-to-Image Generation with Human Priors
Oran Gafni, Adam Polyak, Oron Ashual, Shelly Sheynin, Devi Parikh, Yaniv Taigman
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 scene tokens (ablation) | base transformer · MS-COCO val | FID: 18.01 -> 19.16 | Hurt |
| add scene tokens (human preference) (ablation) | base transformer · MS-COCO val | human preference: 57.3% image quality, 65.3% photorealism, 58.3% text alignment | Improved |
| add face-aware VQ loss (ablation) | +Scene tokens · MS-COCO val | FID: 19.16 -> 14.45 | Improved |
| add classifier-free guidance (ablation) | +Face-aware · MS-COCO val | FID: 14.45 -> 7.55 | Improved |
| add object-aware VQ at 512x512 (ablation) | +CF · MS-COCO val | FID: 7.55 -> 8.70 | Hurt |
| add object-aware VQ at 512x512 (human preference) (ablation) | +CF · MS-COCO val | human preference: 62.0% quality, 53.5% photorealism, 52.2% text alignment | Mixed |
| give scene as additional input (ablation) | Make-A-Scene with CF · MS-COCO val | FID: 7.55 -> 4.69 | Improved |
| use Make-A-Scene instead of LAFITE | text-to-image · MS-COCO val | FID: 8.12 -> 7.55 | Improved |
| use Make-A-Scene instead of GLIDE (filtered) | text-to-image · MS-COCO val (filtered) | FID: 12.24 -> 11.84 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 5 of 5 on most shared results.
- CogView 256 — CogView: Mastering Text-to-Image Generation via Transformers (NeurIPS 2021) · better on 2 of 2 shared results
- DF-GAN — DF-GAN: A Simple and Effective Baseline for Text-to-Image Synthesis (CVPR 2022) · better on 1 of 1 shared result
- GLIDE — GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models (ICML 2022) · better on 1 of 1 shared result
- DM-GAN — DM-GAN: Dynamic Memory Generative Adversarial Networks for Text-To-Image Synthesis (CVPR 2019) · better on 1 of 1 shared result
- AttnGAN — AttnGAN: Fine-Grained Text to Image Generation With Attentional Generative Adversarial Networks (CVPR 2018) · better on 1 of 1 shared result
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
- LLM Blueprint: Enabling Text-to-Image Generation with Complex and Detailed Prompts (ICLR 2024 · similarity 0.90)
- Rich Human Feedback for Text-to-Image Generation (CVPR 2024 · similarity 0.89)
- TextCenGen: Attention-Guided Text-Centric Background Adaptation for Text-to-Image Generation (ICML 2025 · similarity 0.88)
- PreciseCam: Precise Camera Control for Text-to-Image Generation (CVPR 2025 · similarity 0.88)
- Brush Your Text: Synthesize Any Scene Text on Images via Diffusion Model (AAAI 2024 · similarity 0.88)
- DreamText: High Fidelity Scene Text Synthesis (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.
