Research map / 3D vision / 3D generation
ECHO: Ego-centric Modeling of Human-Object Interactions
Ilya A. Petrov, Vladimir Guzov, Riccardo Marin, Emre Aksan, Xu Chen, Daniel Cremers, Thabo Beeler, 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 |
|---|---|---|---|
| use tri-variate diffusion (ECHO) instead of BoDiffusion-HOI baseline | ECHO · BEHAVE | Object E_v2v (cm): 44.2 -> 33.5 | Improved |
| use ECHO instead of EgoAllo-HOI | ECHO · BEHAVE | Object E_v2v (cm): 39.1 -> 33.5 | Improved |
| ECHO vs EgoAllo-HOI foot contact | ECHO · OMOMO | Foot contact: 0.95 -> 0.93 | Hurt |
| add contact modality to the diffusion (three modalities) (ablation) | ECHO · BEHAVE | MPJPE (cm): 8.1 -> 6.8 | Improved |
| train jointly with AMASS (ablation) | ECHO · BEHAVE | MPJPE (cm): 8.7 -> 6.8 | Improved |
| train jointly with AMASS (ablation) | ECHO · AMASS | MPJPE (cm): 43.1 -> 7.4 | Improved |
| joint training instead of two-stage AMASS pretrain then fine-tune (ablation) | ECHO · AMASS | MPJPE (cm): 15.9 -> 7.4 | Improved |
| add smooth inpainting versus standard inpainting (ablation) | ECHO · BEHAVE | MPJPE (cm): 6.9 -> 6.8 | Improved |
| add inference-time contact guidance (ablation) | ECHO · BEHAVE | Object E_v2v (cm): 33.6 -> 33.5 | No clear effect |
| drop 25-50% of hand tracking | ECHO · BEHAVE | Object E_v2v (cm): 33.5 -> 33.9 | No clear effect |
| drop 90% of hand tracking | ECHO · BEHAVE | Object E_v2v (cm): 33.5 -> 36.8 | Hurt |
| add sparse object tracking (50% of frames) | ECHO · BEHAVE | Object E_v2v (cm): 33.46 -> 10.75 | Improved |
Closest papers by meaning
- CG-HOI: Contact-Guided 3D Human-Object Interaction Generation (CVPR 2024 · similarity 0.87)
- HOI-M^3: Capture Multiple Humans and Objects Interaction within Contextual Environment (CVPR 2024 · similarity 0.87)
- Estimating Ego-Body Pose from Doubly Sparse Egocentric Video Data (NeurIPS 2024 · similarity 0.86)
- DiffGrasp: Whole-Body Grasping Synthesis Guided by Object Motion Using a Diffusion Model (AAAI 2025 · similarity 0.86)
- Capturing Closely Interacted Two-Person Motions with Reaction Priors (CVPR 2024 · similarity 0.86)
- CoDA: Coordinated Diffusion Noise Optimization for Whole-Body Manipulation of Articulated Objects (NeurIPS 2025 · 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.
