Research map / Video / Video generation
AnyID: Ultra-Fidelity Universal Identity-Preserving Video Generation from Any Visual References
Jiahao Wang, Hualian Sheng, Sijia Cai, Yuxiao Yang, Weizhan Zhang, Caixia Yan, Bing Deng, Jieping Ye
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 primary reference and differential prompt (ablation) | AnyID-5B · AnyID benchmark | Ele-CLIP: 63.87 -> 68.78 | Improved |
| add primary reference and differential prompt (ablation) | AnyID-5B · AnyID benchmark | Ele-DINO: 64.03 -> 69.50 | Improved |
| add primary reference and differential prompt (ablation) | AnyID-5B · AnyID benchmark | App.: 80.82 -> 86.56 | Improved |
| add primary reference and differential prompt (ablation) | AnyID-5B · AnyID benchmark | Holi-Arc: 72.58 -> 73.22 | Improved |
| add RL (DPO) tuning on human preferences (ablation) | AnyID-5B · AnyID benchmark | Holi-Arc: 72.61 -> 73.22 | Improved |
| add RL (DPO) tuning on human preferences (ablation) | AnyID-5B · AnyID benchmark | Ele-CLIP: 65.49 -> 68.78 | Improved |
| add RL (DPO) tuning on human preferences (ablation) | AnyID-5B · AnyID benchmark | App.: 80.77 -> 86.56 | Improved |
| add RL (DPO) tuning on human preferences (ablation) | AnyID-5B · AnyID benchmark | Mot.: 57.80 -> 60.31 | Improved |
| add auxiliary references at inference (ablation) | AnyID-5B · AnyID benchmark | Holi-Arc: 72.38 -> 73.22 | Improved |
| add auxiliary references at inference (ablation) | AnyID-5B · AnyID benchmark | Holi-Cur: 66.77 -> 67.52 | Improved |
| add auxiliary references at inference (ablation) | AnyID-5B · AnyID benchmark | Ele-CLIP: 68.61 -> 68.78 | No clear effect |
| add auxiliary references at inference (ablation) | AnyID-5B · AnyID benchmark | Ele-DINO: 69.55 -> 69.50 | No clear effect |
Closest papers by meaning
- ARGUS: Stacked Multi-View Identity Mosaic Injection for Subject-Preserving Video Generation (NeurIPS 2026 · similarity 0.87)
- Identity-Preserving Text-to-Video Generation by Frequency Decomposition (CVPR 2025 · similarity 0.87)
- FaithfulFaces: Pose-Faithful Facial Identity Preservation for Text-to-Video Generation (NeurIPS 2026 · similarity 0.87)
- StableAnimator: High-Quality Identity-Preserving Human Image Animation (CVPR 2025 · similarity 0.87)
- Advancing Narrative Long Video Generation via Training-Free Identity-Aware Memory (NeurIPS 2026 · similarity 0.86)
- Animate Anyone: Consistent and Controllable Image-to-Video Synthesis for Character Animation (CVPR 2024 · similarity 0.86)
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.
