Research map / 3D vision / 3D generation
See and Think: Disentangling Semantic Scene Completion
Shice Liu, YU HU, Yiming Zeng, Qiankun Tang, Beibei Jin, Yinhe Han, Xiaowei Li
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 explicit 2D semantic segmentation network (SNet) before 3D completion (ablation) | SATNet · NYUv2 | SSC mean IoU: 29.0 -> 33.1 | Improved |
| add explicit SNet (ablation) | SATNet · SUNCG-D | SSC mean IoU: 52.3 -> 64.3 | Improved |
| initialize SNet with segmentation-pretrained weights (ablation) | SATNet Depth · SUNCG-D | SSC mean IoU: 53.5 -> 64.3 | Improved |
| initialize SNet with segmentation-pretrained weights (ablation) | SATNet Depth · SUNCG-RGBD | SSC mean IoU: 34.5 -> 43.4 | Improved |
| initialize SNet with segmentation-pretrained weights (ablation) | SATNet RGBD · SUNCG-RGBD | SSC mean IoU: 37.8 -> 41.2 | Improved |
| add ASPP to TNet (ablation) | SATNet Depth · SUNCG-D | SSC mean IoU: 61.9 -> 64.3 | Improved |
| add ASPP to TNet (ablation) | SATNet Depth · SUNCG-RGBD | SSC mean IoU: 41.5 -> 43.4 | Improved |
| add ASPP to TNet (ablation) | SATNet RGBD · SUNCG-RGBD | SSC mean IoU: 39.7 -> 41.2 | Improved |
| replace estimated segmentation with ground-truth segmentation (ablation) | SATNet · NYUv2 | SSC mean IoU: 33.1 -> 40.8 | Improved |
| fuse RGB and depth at end of TNet instead of single-branch depth | SATNet · NYUv2 | SSC mean IoU: 33.1 -> 34.4 | Improved |
| use single-branch RGB-D input instead of depth only | SATNet · NYUv2 | SSC mean IoU: 33.1 -> 33.0 | No clear effect |
| TNetFuse instead of SNetFuse | SATNet · NYUv2 | SSC mean IoU: 33.6 -> 34.4 | Improved |
Closest papers by meaning
- Semi-supervised 3D Semantic Scene Completion with 2D Vision Foundation Model Guidance (AAAI 2025 · similarity 0.91)
- Disentangling Instance and Scene Contexts for 3D Semantic Scene Completion (ICCV 2025 · similarity 0.91)
- MonoScene: Monocular 3D Semantic Scene Completion (CVPR 2022 · similarity 0.91)
- VisHall3D: Monocular Semantic Scene Completion from Reconstructing the Visible Regions to Hallucinating the Invisible Regions (ICCV 2025 · similarity 0.90)
- Bi-SSC: Geometric-Semantic Bidirectional Fusion for Camera-based 3D Semantic Scene Completion (CVPR 2024 · similarity 0.90)
- VLScene: Vision-Language Guidance Distillation for Camera-Based 3D Semantic Scene Completion (AAAI 2025 · similarity 0.89)
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.
