Research map / 3D vision / Face and shape
CASA: Category-agnostic Skeletal Animal Reconstruction
Yuefan Wu, Zeyuan Chen, Shaowei Liu, Zhongzheng Ren, Shenlong Wang
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 mask consistency energy (ablation) | CASA · PlanetZoo | mIoU: 0.011 -> 0.512 | Improved |
| add optical flow energy (ablation) | CASA · PlanetZoo | mIoU: 0.427 -> 0.512 | Improved |
| add symmetry offset energy (ablation) | CASA · PlanetZoo | mIoU: 0.387 -> 0.512 | Improved |
| add motion smoothness energy (ablation) | CASA · PlanetZoo | mIoU: 0.449 -> 0.512 | Improved |
| add motion smoothness energy (ablation) | CASA · PlanetZoo | mCham: 0.041 -> 0.053 | Hurt |
| add neural offset field instead of no offset (ablation) | CASA · PlanetZoo | mIoU: 0.144 -> 0.512 | Improved |
| use neural displacement field instead of per-vertex offset (ablation) | CASA · PlanetZoo | mIoU: 0.235 -> 0.512 | Improved |
| add stretchable skeleton (ablation) | CASA · PlanetZoo | mIoU: 0.372 -> 0.512 | Improved |
| add shape scaling step (ablation) | CASA · PlanetZoo | mIoU: 0.299 -> 0.512 | Improved |
| use CLIP instead of ImageNet features for retrieval (ablation) | CASA · PlanetZoo | mIoU: 0.111 -> 0.217 | Improved |
| add inverse-graphics optimization after retrieval (ablation) | CASA · PlanetZoo | mIoU: 0.217 -> 0.512 | Improved |
| initialize rigging from retrieved template instead of k-means (ablation) | CASA · PlanetZoo | mIoU: 0.305 -> 0.512 | 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.
- LASR — LASR: Learning Articulated Shape Reconstruction From a Monocular Video (CVPR 2021) · better on 4 of 5 shared results
- ViSER — ViSER: Video-Specific Surface Embeddings for Articulated 3D Shape Reconstruction (NeurIPS 2021) · better on 5 of 5 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
- Hierarchically Structured Neural Bones for Reconstructing Animatable Objects from Casual Videos (ECCV 2024 · similarity 0.89)
- Learning Implicit Representation for Reconstructing Articulated Objects (ICLR 2024 · similarity 0.88)
- REACTO: Reconstructing Articulated Objects from a Single Video (CVPR 2024 · similarity 0.87)
- SAOR: Single-View Articulated Object Reconstruction (CVPR 2024 · similarity 0.86)
- S3O: A Dual-Phase Approach for Reconstructing Dynamic Shape and Skeleton of Articulated Objects from Single Monocular Video (ICML 2024 · similarity 0.86)
- SAMoR: Motion Modelling for Articulated Objects of Any Skeleton and Topology (NeurIPS 2026 · similarity 0.85)
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.
