Research map / Video / Video generation
WildActor: Unconstrained Identity-Preserving Video Generation
Qin Guo, Tianyu Yang, Xuanhua He, Fei Shen, Yong Zhang, Zhuoliang Kang, Xiaoming Wei, Dan Xu
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 |
|---|---|---|---|
| use Actor-18M with random sampling instead of raw-crop references (ablation) | WildActor · Ablation (data strategies) | Body Consistency average: 0.802 -> 0.865 | Improved |
| add viewpoint-adaptive sampling (ablation) | WildActor · Ablation (data strategies) | Body Consistency average: 0.865 -> 0.952 | Improved |
| add viewpoint-adaptive sampling (ablation) | WildActor · Ablation (data strategies) | Body Consistency (Back): 0.785 -> 0.937 | Improved |
| add Actor-18M with viewpoint-adaptive sampling over raw-crop (ablation) | WildActor · Ablation (data strategies) | Body Consistency (Front): 0.885 -> 0.958 | Improved |
| add AIPA (vs full self-attention) (ablation) | WildActor · Ablation (model components) | Face ID: 0.515 -> 0.559 | Improved |
| add AIPA (vs full self-attention) (ablation) | WildActor · Ablation (model components) | Semantic Alignment (VLM-level): 0.610 -> 0.920 | Improved |
| add AIPA (vs full self-attention) (ablation) | WildActor · Ablation (model components) | Body Consistency: 0.890 -> 0.952 | Improved |
| add I-RoPE (ablation) | WildActor · Ablation (model components) | Body Consistency: 0.825 -> 0.952 | Improved |
| add I-RoPE (ablation) | WildActor · Ablation (model components) | Face ID: 0.542 -> 0.559 | Improved |
| add identity references in autoregressive T2V->I2V | T2V -> I2V · Sequential Narrative | Face Identity: 0.320 -> 0.548 | Improved |
| add identity references in autoregressive T2V->I2V | T2V -> I2V · Sequential Narrative | Body Consistency: 0.450 -> 0.925 | Improved |
Closest papers by meaning
- FaithfulFaces: Pose-Faithful Facial Identity Preservation for Text-to-Video Generation (NeurIPS 2026 · similarity 0.90)
- HumanVid: Demystifying Training Data for Camera-controllable Human Image Animation (NeurIPS 2024 · similarity 0.87)
- Vid2Avatar-Pro: Authentic Avatar from Videos in the Wild via Universal Prior (CVPR 2025 · similarity 0.87)
- ReGenHuman: Re-Generating Human Appearances for Realistic Full-Body Video Anonymization (NeurIPS 2026 · similarity 0.87)
- ARGUS: Stacked Multi-View Identity Mosaic Injection for Subject-Preserving Video Generation (NeurIPS 2026 · similarity 0.87)
- StableAnimator: High-Quality Identity-Preserving Human Image Animation (CVPR 2025 · similarity 0.87)
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.
