Research map / 3D vision / 3D generation
Holistic 3D Scene Understanding From a Single Image With Implicit Representation
Cheng Zhang, Zhaopeng Cui, Yinda Zhang, Bing Zeng, Marc Pollefeys, Shuaicheng Liu
Check 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 image-based local implicit shape representation | Object reconstruction · Pix3D | Chamfer distance (×10^3): 8.36 -> 6.72 | Improved |
| Add scene graph convolution to Total3D | Total3D · SUN RGB-D | Detection mAP: 26.38 -> 37.04 | Improved |
| Add scene graph convolution to Total3D | Total3D · SUN RGB-D | Layout IoU: 59.25 -> 62.49 | Improved |
| Add GCN refinement (ablation) | Holistic scene understanding · SUN RGB-D | Detection mAP: 27.47 -> 45.21 | Improved |
| Add GCN refinement (ablation) | Holistic scene understanding · SUN RGB-D | Layout IoU: 60.04 -> 64.41 | Improved |
| Include element centers in GCN node features (ablation) | Holistic scene understanding · SUN RGB-D | Detection mAP: 42.05 -> 45.21 | Improved |
| Include analytic code in GCN node features (ablation) | Holistic scene understanding · SUN RGB-D | Detection mAP: 43.10 -> 45.21 | Improved |
| Add physical violation loss (ablation) | Holistic scene understanding · SUN RGB-D | Detection mAP: 43.33 -> 45.21 | Improved |
| Add physical violation loss (ablation) | Holistic scene understanding · SUN RGB-D | Collision volume (dm³/scene): 16.64 -> 13.55 | Improved |
| Add physical violation loss (ablation) | Holistic scene understanding · SUN RGB-D | Supporting error (cm): 13.35 -> 14.71 | Hurt |
| Add physical violation loss (ablation) | Holistic scene understanding · SUN RGB-D | Layout IoU: 63.52 -> 64.41 | Improved |
| Use full method rather than Total3D | 3D layout estimation · SUN RGB-D | Layout IoU: 59.2 -> 64.4 | Improved |
Closest papers by meaning
- DeepContext: Context-Encoding Neural Pathways for 3D Holistic Scene Understanding (ICCV 2017 · similarity 0.88)
- Learning 3D Scene Priors With 2D Supervision (CVPR 2023 · similarity 0.88)
- Holistic++ Scene Understanding: Single-View 3D Holistic Scene Parsing and Human Pose Estimation With Human-Object Interaction and Physical Commonsense (ICCV 2019 · similarity 0.88)
- Learning 3D Object Shape and Layout Without 3D Supervision (CVPR 2022 · similarity 0.88)
- ImageNet3D: Towards General-Purpose Object-Level 3D Understanding (NeurIPS 2024 · similarity 0.87)
- Implicit Zoo: A Large-Scale Dataset of Neural Implicit Functions for 2D Images and 3D Scenes (NeurIPS 2024 · 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.
