Research map / 3D vision / 3D generation
Perceiving 3D Human-Object Spatial Arrangements from a Single Image in the Wild
Jason Y. Zhang, Sam Pepose, Hanbyul Joo, Deva Ramanan, Jitendra Malik, Angjoo Kanazawa
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 |
|---|---|---|---|
| reason jointly about humans and objects instead of independent composition | PHOSA · COCO-2017 test subset (average) | % images where full method preferred: 75.0 | Improved |
| add occlusion-aware silhouette loss in the joint optimization (ablation) | PHOSA · COCO-2017 test subset (average) | % images where full method preferred: 79.9 | Improved |
| add interaction loss (ablation) | PHOSA · COCO-2017 test subset (average) | % images where full method preferred: 66.3 | Improved |
| add scale loss (ablation) | PHOSA · COCO-2017 test subset (average) | % images where full method preferred: 56.4 | Improved |
| add ordinal depth loss (ablation) | PHOSA · COCO-2017 test subset (average) | % images where full method preferred: 52.3 | Improved |
| add collision loss (ablation) | PHOSA · COCO-2017 test subset (average) | % images where full method preferred: 49.5 | No clear effect |
| add interaction loss (ablation) | PHOSA · COCO-2017 test subset (Laptop) | % images where full method preferred: 46 | Hurt |
| add scale loss (ablation) | PHOSA · COCO-2017 test subset (Bench) | % images where full method preferred: 49 | No clear effect |
| add collision loss (ablation) | PHOSA · COCO-2017 test subset (Bench) | % images where full method preferred: 40 | Hurt |
| add occlusion-aware silhouette loss (ablation) | PHOSA · COCO-2017 test subset (Skateboard) | % images where full method preferred: 96 | Improved |
Closest papers by meaning
- Human-Aware Object Placement for Visual Environment Reconstruction (CVPR 2022 · similarity 0.90)
- ODHSR: Online Dense 3D Reconstruction of Humans and Scenes from Monocular Videos (CVPR 2025 · similarity 0.88)
- Reconstructing In-the-Wild Open-Vocabulary Human-Object Interactions (CVPR 2025 · similarity 0.88)
- Reconstructing People, Places, and Cameras (CVPR 2025 · similarity 0.88)
- PICO: Reconstructing 3D People In Contact with Objects (CVPR 2025 · similarity 0.88)
- Evaluating Multiview Object Consistency in Humans and Image Models (NeurIPS 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.
