Research map / 3D vision / 3D generation
Root Pose Decomposition Towards Generic Non-rigid 3D Reconstruction with Monocular Videos
Yikai Wang, Yinpeng Dong, Fuchun Sun, Xiao Yang
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 RPD to BANMo when CSE/real pose is unavailable (ablation) | BANMo · AMA-swing (multi-video, no CSE pose) | F@2%: 12.5 -> 53.2 | Improved |
| add RPD to ViSER when CSE pose is unavailable (ablation) | ViSER · AMA-swing (multi-video, no CSE pose) | F@2%: 10.3 -> 51.5 | Improved |
| add RPD to BANMo when CSE pose is available (ablation) | BANMo · AMA-swing (multi-video, CSE pose) | F@2%: 57.0 -> 58.2 | Improved |
| remove CSE pose from BANMo (ablation) | BANMo · DAVIS (Human) | mIoU (%): 72.7 -> 35.1 | Hurt |
| use RPD without CSE pose instead of BANMo | OVIS (Cats) · OVIS (Cats) | mIoU (%): 53.6 -> 87.7 | Improved |
| increase initial pose error to 90 degrees (ablation) | RPD · AMA-swing | Chamfer distance (cm): 9.6 -> 10.3 | No clear effect |
| increase initial pose error to 135 degrees (ablation) | RPD · AMA-swing | Chamfer distance (cm): 9.6 -> 16.4 | Hurt |
| add CSE pose initialization to RPD (ablation) | RPD · AMA-swing (multi-video) | F@2%: 53.9 -> 58.7 | Improved |
| add root pose decomposition (RPD) during training | reconstruction pipeline · human reconstruction | convergence / pose ambiguity | Improved |
| increase number of pyramid levels in pose network | RPD · qualitative reconstruction | reconstruction detail | Improved |
| use 30 instead of 25 control points | RPD · OVIS (fish) | reconstruction quality | No clear effect |
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.
- BANMo — BANMo: Building Animatable 3D Neural Models From Many Casual Videos (CVPR 2022) · better on 9 of 9 shared results
- ViSER — ViSER: Video-Specific Surface Embeddings for Articulated 3D Shape Reconstruction (NeurIPS 2021) · better on 1 of 1 shared result
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
- RGBD Objects in the Wild: Scaling Real-World 3D Object Learning from RGB-D Videos (CVPR 2024 · similarity 0.87)
- Free-Moving Object Reconstruction and Pose Estimation with Virtual Camera (AAAI 2025 · similarity 0.86)
- Hand-held Object Reconstruction from RGB Video with Dynamic Interaction (CVPR 2025 · similarity 0.86)
- ODHSR: Online Dense 3D Reconstruction of Humans and Scenes from Monocular Videos (CVPR 2025 · similarity 0.86)
- SLAM3R: Real-Time Dense Scene Reconstruction from Monocular RGB Videos (CVPR 2025 · similarity 0.85)
- Unsupervised Learning of Category-Level 3D Pose from Object-Centric Videos (CVPR 2024 · 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.
