Research map / 3D vision / Face and shape
MagicPony: Learning Articulated 3D Animals in the Wild
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 the Eikonal regularizer (ablation) | MagicPony · Toy Bird Scans | Chamfer Distance (cm): 1.19 -> 0.79 | Helped |
| add the deformation regularizer (ablation) | MagicPony · Toy Bird Scans | Chamfer Distance (cm): 1.37 -> 0.79 | Helped |
| use the viewpoint hypothesis temperature schedule (ablation) | MagicPony · Toy Bird Scans | Chamfer Distance (cm): 1.66 -> 0.79 | Helped |
| add the bilateral symmetry constraint (ablation) | MagicPony · Toy Bird Scans | Chamfer Distance (cm): 1.56 -> 0.79 | Helped |
| add articulation while retaining symmetry (ablation) | MagicPony · Toy Bird Scans | Chamfer Distance (cm): 1.63 -> 0.79 | Helped |
| add the image reconstruction loss (ablation) | MagicPony · Toy Bird Scans | Chamfer Distance (cm): 2.11 -> 0.79 | Helped |
| use the full model's bone configuration rather than 4 bones (ablation) | MagicPony · Toy Bird Scans | Chamfer Distance (cm): 2.10 -> 0.79 | Helped |
| use the full model's bone configuration rather than 16 bones (ablation) | MagicPony · Toy Bird Scans | Chamfer Distance (cm): 1.27 -> 0.79 | Helped |
| add DINO feature rendering loss (ablation) | MagicPony · Animal-category training images | viewpoint coverage | Helped |
| use multi-hypothesis viewpoint prediction (ablation) | MagicPony · Animal-category training images | viewpoint coverage | Helped |
| combine feature rendering loss and multi-hypothesis viewpoint prediction (ablation) | MagicPony · Toy Bird Scans | Chamfer Distance (cm) | Helped |
| finetune the albedo network on the test image | MagicPony · Out-of-distribution textures, including drawings | texture quality: 100 iterations | Helped |
Closest papers by meaning
- LASSIE: Learning Articulated Shapes from Sparse Image Ensemble via 3D Part Discovery (NeurIPS 2022 · similarity 0.91)
- SAOR: Single-View Articulated Object Reconstruction (CVPR 2024 · similarity 0.90)
- Implicit Zoo: A Large-Scale Dataset of Neural Implicit Functions for 2D Images and 3D Scenes (NeurIPS 2024 · similarity 0.87)
- VAREN: Very Accurate and Realistic Equine Network (CVPR 2024 · similarity 0.87)
- Learning the 3D Fauna of the Web (CVPR 2024 · similarity 0.86)
- BEAST3D: animal behavioral analysis and neural encoding from multi-view video via Gaussian splatting (NeurIPS 2026 · 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.