Research map / Video / Video generation
MagicID: Hybrid Preference Optimization for ID-Consistent and Dynamic-Preserved Video Customization
Hengjia Li, Lifan Jiang, Xi Xiao, Tianyang Wang, Hongwei Yi, Boxi Wu, Deng Cai
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 dynamic-preferred pairs to ID-preferred pairs (ablation) | MagicID · hybrid pair selection | Dynamic Degree: 7.382 -> 14.42 | Improved |
| add dynamic-preferred pairs to ID-preferred pairs (ablation) | MagicID · hybrid pair selection | Face Sim.: 0.605 -> 0.600 | No clear effect |
| add dynamic-preferred pairs to ID-preferred pairs (ablation) | MagicID · hybrid pair selection | CLIP-T: 25.94 -> 26.28 | Improved |
| add ID-preferred pairs preference optimization instead of self-reconstruction (ablation) | DreamBooth-style baseline · hybrid pair selection | Face Sim.: 0.276 -> 0.605 | Improved |
| add dynamic reward to ID reward (ablation) | MagicID · customized reward | Dynamic Degree: 6.332 -> 12.33 | Improved |
| add dynamic reward to ID reward (ablation) | MagicID · customized reward | Face Sim.: 0.598 -> 0.607 | No clear effect |
| add semantic reward to ID + dynamic reward (ablation) | MagicID · customized reward | Dynamic Degree: 12.33 -> 14.42 | Improved |
| add semantic reward to ID + dynamic reward (ablation) | MagicID · customized reward | CLIP-T: 25.73 -> 26.28 | Improved |
| add semantic reward to ID + dynamic reward (ablation) | MagicID · customized reward | Face Sim.: 0.607 -> 0.600 | Hurt |
Closest papers by meaning
- Identity-Preserving Text-to-Video Generation by Frequency Decomposition (CVPR 2025 · similarity 0.86)
- VideoMage: Multi-Subject and Motion Customization of Text-to-Video Diffusion Models (CVPR 2025 · similarity 0.85)
- ARGUS: Stacked Multi-View Identity Mosaic Injection for Subject-Preserving Video Generation (NeurIPS 2026 · similarity 0.85)
- StableAnimator: High-Quality Identity-Preserving Human Image Animation (CVPR 2025 · similarity 0.84)
- FaithfulFaces: Pose-Faithful Facial Identity Preservation for Text-to-Video Generation (NeurIPS 2026 · similarity 0.84)
- CustomCrafter: Customized Video Generation with Preserving Motion and Concept Composition Abilities (AAAI 2025 · 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.
