Research map / Image generation / Text-to-image diffusion
AddMe: Zero-shot Group-photo Synthesis by Inserting People into Scenes
Dongxu Yue, Maomao Li, Yunfei Liu, Ailing Zeng, Tianyu Yang, Qin Guo, Yu Li
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 disentangled ID-Adapter to naive IP-Adapter inpainting (ablation) | naive IP-Adapter inpainting · AddMe benchmark | Face Sim. (%): 29.03 -> 56.79 | Improved |
| add disentangled ID-Adapter (ablation) | naive IP-Adapter inpainting · AddMe benchmark | FID: 34.87 -> 35.45 | Hurt |
| add Enhanced Portrait Attention module (ablation) | ID-Adapter · AddMe benchmark | FID: 35.45 -> 27.16 | Improved |
| add Enhanced Portrait Attention module (ablation) | ID-Adapter · AddMe benchmark | QS: 6.68 -> 11.13 | Improved |
| add data augmentation of reference ID image (ablation) | AddMe · AddMe benchmark | FID: 27.16 -> 26.11 | Improved |
| add data augmentation of reference ID image (ablation) | AddMe · AddMe benchmark | Face Sim. (%): 59.73 -> 60.15 | Improved |
| replace AnyDoor with AddMe | exemplar-based editing · AddMe benchmark | Face Sim. (%): 14.21 -> 60.15 | Improved |
| replace UniPaint test-time fine-tuning with AddMe tuning-free | UniPaint · AddMe benchmark | editing time: 64.93s -> 2.44s | Improved |
| use AddMe instead of PbE for image quality | PbE · AddMe benchmark | QS: 14.83 -> 12.09 | Hurt |
Closest papers by meaning
- PAIR Diffusion: A Comprehensive Multimodal Object-Level Image Editor (CVPR 2024 · similarity 0.86)
- Diffusion Self-Distillation for Zero-Shot Customized Image Generation (CVPR 2025 · similarity 0.85)
- EchoShot: Multi-Shot Portrait Video Generation (NeurIPS 2025 · similarity 0.85)
- HyperLoRA: Parameter-Efficient Adaptive Generation for Portrait Synthesis (CVPR 2025 · similarity 0.85)
- DiffPortrait3D: Controllable Diffusion for Zero-Shot Portrait View Synthesis (CVPR 2024 · similarity 0.85)
- When StyleGAN Meets Stable Diffusion: a W+ Adapter for Personalized Image Generation (CVPR 2024 · similarity 0.85)
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.
