Research map / Video / Video generation
Kirin: Animal Motion Generation from In-the-Wild Video
Brian Nlong Zhao, Zhuoyang Pan, James Rehg, Jiajun Wu, Elliott (Shangzhe) Wu
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 image conditioning to text-conditioned model (ablation) | Kirin · AiM3D test set | FID: 11.889 -> 6.248 | Improved |
| add image conditioning to text-conditioned model (ablation) | Kirin · AiM3D test set | R-Precision Top-1: 0.032 -> 0.043 | Improved |
| add image conditioning to text-conditioned model (ablation) | Kirin · AiM3D test set | Diversity: 4.693 -> 4.319 | Hurt |
| retrain AniMo on AiM3D data instead of AniMo4D | AniMo · AiM3D test set | FID: 30.043 -> 30.516 | No clear effect |
| use Kirin architecture instead of AniMo on same data | Kirin (text) · AiM3D test set | FID: 30.516 -> 11.889 | Improved |
| train on AiM3D with Kirin, test on unseen AnimalML3D | Kirin · AnimalML3D | FID: 150.671 -> 138.145 | Improved |
| image-only conditioning versus text+image (ablation) | Kirin · AiM3D test set | FID: 6.248 -> 25.089 | Hurt |
| sequence-level refinement with global tracking over AniMer | Kirin reconstruction · 4D animal reconstruction benchmark | PCK@0.1: 0.537 -> 0.751 | Improved |
Closest papers by meaning
- OmniMotionGPT: Animal Motion Generation with Limited Data (CVPR 2024 · similarity 0.90)
- AniMo: Species-Aware Model for Text-Driven Animal Motion Generation (CVPR 2025 · similarity 0.88)
- How to Move Your Dragon: Text-to-Motion Synthesis for Large-Vocabulary Objects (ICML 2025 · similarity 0.88)
- Web-Scale Collection of Video Data for 4D Animal Reconstruction (NeurIPS 2025 · similarity 0.87)
- BEAST3D: animal behavioral analysis and neural encoding from multi-view video via Gaussian splatting (NeurIPS 2026 · similarity 0.87)
- Animal-Bench: Benchmarking Multimodal Video Models for Animal-centric Video Understanding (NeurIPS 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.
