Research map / 3D vision / 3D generation
TriDi: Trilateral Diffusion of 3D Humans, Objects, and Interactions
Ilya A. Petrov, Riccardo Marin, Julian Chibane, Gerard Pons-Moll
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 left-right mirror augmentation (ablation) | TriDi · BEHAVE, human generation given object | 1-NNA: 69.74 -> 67.89 | Improved |
| add mirror augmentation (ablation) | TriDi · BEHAVE, human generation given object | MPJPE: 23.2 -> 20.8 | Improved |
| add mirror augmentation (ablation) | TriDi · BEHAVE, object generation given human | E_v2v: 31.0 -> 28.0 | Improved |
| add mirror augmentation (ablation) | TriDi · GRAB, human generation given object | MPJPE: 14.1 -> 15.3 | Hurt |
| add contact guidance during denoising (ablation) | TriDi · BEHAVE, human generation given object | MPJPE: 21.5 -> 20.8 | Improved |
| add contact guidance during denoising (ablation) | TriDi · GRAB, object generation given human | E_v2v: 9.0 -> 6.9 | Improved |
| add interaction modality to diffusion (three-way instead of human-object) (ablation) | TriDi · BEHAVE, human generation given object | MPJPE: 21.9 -> 20.8 | Improved |
| add interaction modality to diffusion (ablation) | TriDi · GRAB, object generation given human | E_v2v: 9.5 -> 6.9 | Improved |
| add interaction modality to diffusion (ablation) | TriDi · BEHAVE, object generation given human | 1-NNA: 65.18 -> 63.72 | Improved |
| train one joint model instead of single-mode model | TriDi · BEHAVE, human generation given object | MPJPE: 21.0 -> 20.8 | No clear effect |
| use TriDi instead of GNet | human generation · BEHAVE, human generation given object | MMD: 1.789 -> 1.352 | Improved |
| use TriDi instead of ObjPOP+cVAE | object generation · GRAB, object generation given human | MMD: 0.483 -> 0.268 | Improved |
Closest papers by meaning
- HOI-M^3: Capture Multiple Humans and Objects Interaction within Contextual Environment (CVPR 2024 · similarity 0.91)
- InterAct: Advancing Large-Scale Versatile 3D Human-Object Interaction Generation (CVPR 2025 · similarity 0.91)
- CG-HOI: Contact-Guided 3D Human-Object Interaction Generation (CVPR 2024 · similarity 0.91)
- Reconstructing In-the-Wild Open-Vocabulary Human-Object Interactions (CVPR 2025 · similarity 0.90)
- InteractAnything: Zero-shot Human Object Interaction Synthesis via LLM Feedback and Object Affordance Parsing (CVPR 2025 · similarity 0.90)
- PICO: Reconstructing 3D People In Contact with Objects (CVPR 2025 · similarity 0.90)
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.
