Research map / 3D vision / Face and shape
Ponymation: Learning Articulated 3D Animal Motions from Unlabeled Online Videos
Keqiang Sun, Dor Litvak, Yunzhi Zhang, Hongsheng Li, Jiajun Wu, 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 |
|---|---|---|---|
| train from videos rather than single images (Ponymation vs MagicPony) | MagicPony · PASCAL VOC horses | PCK: 42.8% -> 48.0% | Improved |
| train from videos rather than single images | MagicPony · APT-36K horses | Velocity Error: 57.3% -> 49.1% | Improved |
| add AnimalMotion video data to MagicPony + VAE (ablation) | Ponymation · APT-36K motion generation | Motion Chamfer Distance: 38.77 -> 38.12 | Improved |
| add temporal smoothness loss (ablation) | Ponymation · APT-36K motion generation | Motion Chamfer Distance: 38.12 -> 38.03 | No clear effect |
| add spatial transformers in motion VAE encoder and decoder (ablation) | Ponymation · reconstruction | PCK@0.1: 33.4% -> 37.6% | Improved |
| add Teacher Loss for motion VAE training (ablation) | Ponymation · reconstruction | PCK@0.1: 32.4% -> 37.6% | Improved |
| add motion VAE to the reconstruction model (ablation) | Ponymation · reconstruction | PCK@0.1: 44.3% -> 37.6% | Hurt |
| increase training sequence length from 10 to 50 frames (ablation) | Ponymation · APT-36K motion generation | Motion Chamfer Distance: 38.03 -> 39.25 | Hurt |
| set KL loss weight to 0.001 instead of 0.01 (ablation) | Ponymation · reconstruction | PCK@0.1: 33.58% -> 37.63% | Improved |
| use Ponymation instead of text-to-4D 4D-fy | 4D-fy · horse motion generation | Motion Strength: 0.29 -> 4.66 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- MagicPony — MagicPony: Learning Articulated 3D Animals in the Wild (CVPR 2023) · better on 4 of 4 shared results
- 4D-fy — 4D-fy: Text-to-4D Generation Using Hybrid Score Distillation Sampling (CVPR 2024) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Kirin: Animal Motion Generation from In-the-Wild Video (ECCV 2026 · similarity 0.90)
- BEAST3D: animal behavioral analysis and neural encoding from multi-view video via Gaussian splatting (NeurIPS 2026 · similarity 0.88)
- OmniMotionGPT: Animal Motion Generation with Limited Data (CVPR 2024 · similarity 0.88)
- Pose Splatter: A 3D Gaussian Splatting Model for Quantifying Animal Pose and Appearance (NeurIPS 2025 · similarity 0.87)
- Learning the 3D Fauna of the Web (CVPR 2024 · similarity 0.86)
- Web-Scale Collection of Video Data for 4D Animal Reconstruction (NeurIPS 2025 · 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.
