Research map / 3D vision / 3D generation
CARI4D: Category Agnostic 4D Reconstruction of Human-Object Interaction
Xianghui Xie, Bowen Wen, Yan Chang, Hesam Rabeti, Jiefeng Li, Ye Yuan, Gerard Pons-Moll, Stan Birchfield
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 |
|---|---|---|---|
| Use our initialization (pose hypothesis selection + depth-aligned NLF) instead of raw NLF + FoundationPose tracking (ablation) | CARI4D initialization · BEHAVE ablation subset (24 videos) | CD-c (cm): 405.13 -> 10.79 | Improved |
| Use our initialization instead of raw NLF + per-frame FoundationPose (ablation) | CARI4D initialization · BEHAVE ablation subset (24 videos) | CD-c (cm): 13.02 -> 10.79 | Improved |
| Add CoCoNet refinement trained with depth alignment (ablation) | CARI4D initialization · BEHAVE ablation subset (24 videos) | CD-c (cm): 10.79 -> 8.62 | Improved |
| Add depth alignment when training CoCoNet (ablation) | CoCoNet · BEHAVE ablation subset (24 videos) | CD-c (cm): 9.95 -> 8.62 | Improved |
| Add depth alignment when training CoCoNet (ablation) | CoCoNet · BEHAVE ablation subset (24 videos) | CD-h (cm): 8.01 -> 7.01 | Improved |
| Add contact-aware joint optimization after CoCoNet (ablation) | CARI4D · BEHAVE ablation subset (24 videos) | Acc-o: 3.78 -> 0.38 | Improved |
| Add contact-aware joint optimization after CoCoNet (ablation) | CARI4D · BEHAVE ablation subset (24 videos) | Acc-h: 1.75 -> 1.06 | Improved |
| Add contact-aware joint optimization after CoCoNet (ablation) | CARI4D · BEHAVE ablation subset (24 videos) | CD-c (cm): 8.62 -> 9.35 | Mixed |
| Provide ground truth object mesh and depth map as input (ablation) | CARI4D · BEHAVE oracle study | CD-c (cm): 9.35 -> 7.20 | Improved |
| Provide ground truth object mesh only (ablation) | CARI4D · BEHAVE oracle study | CD-o (cm): 11.57 -> 9.26 | Improved |
| Replace VisTracker/InterTrack with CARI4D | baseline trackers · BEHAVE test (62 videos) | CD-c (cm): 14.22 -> 9.23 | Improved |
| Replace VisTracker with CARI4D | VisTracker · BEHAVE test (62 videos) | Acc-h: 0.54 -> 1.14 | Hurt |
Later papers that compared with this method
1 later accepted paper reports CARI4D as a baseline in their results tables; 0 of them report a better number on most of the results they share with it.
- Real2Sim in HOI: Toward Physically Plausible HOI Reconstruction from Monocular Videos (NeurIPS 2026 · better on 3 of 6 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
- Real2Sim in HOI: Toward Physically Plausible HOI Reconstruction from Monocular Videos (NeurIPS 2026 · similarity 0.89)
- UniCon3R: Unified Contact-aware 4D Human-Scene Reconstruction from Monocular Video (NeurIPS 2026 · similarity 0.89)
- Reconstructing In-the-Wild Open-Vocabulary Human-Object Interactions (CVPR 2025 · similarity 0.88)
- Template Free Reconstruction of Human-object Interaction with Procedural Interaction Generation (CVPR 2024 · similarity 0.88)
- BIGS: Bimanual Category-agnostic Interaction Reconstruction from Monocular Videos via 3D Gaussian Splatting (CVPR 2025 · similarity 0.88)
- HOLD: Category-agnostic 3D Reconstruction of Interacting Hands and Objects from Video (CVPR 2024 · similarity 0.88)
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.
