Research map / 3D vision / 3D generation
Semantic Scene Completion via Integrating Instances and Scene In-the-Loop
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 SISNet with DeepLabv3 instead of Sketch | Semantic scene completion · NYU | SSC-mIoU: 41.1 -> 52.4 | Helped |
| use SISNet with DeepLabv3 instead of Sketch | Semantic scene completion · NYUCAD | SSC-mIoU: 55.2 -> 63.5 | Helped |
| use SISNet with DeepLabv3 instead of Sketch | Semantic scene completion · SUNCG-RGBD | SSC-mIoU: 55.6 -> 79.0 | Helped |
| add scene refinement S1 without instance completion (ablation) | SISNet baseline S0 · NYU | SSC-mIoU: 42.5 -> 44.9 | Helped |
| add scene refinement S1 without instance completion (ablation) | SISNet baseline S0 · NYUCAD | SSC-mIoU: 53.6 -> 54.2 | Helped |
| add scene refinement S1 without instance completion (ablation) | SISNet baseline S0 · SUNCG-RGBD | SSC-mIoU: 73.0 -> 74.5 | Helped |
| add two scene-instance-scene iterations instead of one (ablation) | SISNet · NYU | SSC-mIoU: 47.7 -> 49.8 | Helped |
| add two scene-instance-scene iterations instead of one (ablation) | SISNet · NYUCAD | SSC-mIoU: 58.0 -> 59.9 | Helped |
| add two scene-instance-scene iterations instead of one (ablation) | SISNet · SUNCG-RGBD | SSC-mIoU: 76.6 -> 78.8 | Helped |
| remove initial scene completion before instance completion (ablation) | SISNet · NYU | SSC: 47.7 -> 46.6 | Hurt |
| remove initial scene completion before instance completion (ablation) | SISNet · NYUCAD | SSC: 58.0 -> 55.5 | Hurt |
| remove initial scene completion before instance completion (ablation) | SISNet · SUNCG-RGBD | SSC: 76.6 -> 75.8 | Hurt |
Closest papers by meaning
- Semantic Scene Completion From a Single Depth Image (CVPR 2017 · similarity 0.91)
- RevealNet: Seeing Behind Objects in RGB-D Scans (CVPR 2020 · similarity 0.91)
- Context and Geometry Aware Voxel Transformer for Semantic Scene Completion (NeurIPS 2024 · similarity 0.89)
- IPFormer: Visual 3D Panoptic Scene Completion with Context-Adaptive Instance Proposals (NeurIPS 2025 · similarity 0.89)
- Symphonize 3D Semantic Scene Completion with Contextual Instance Queries (CVPR 2024 · similarity 0.88)
- Voxel Proposal Network via Multi-Frame Knowledge Distillation for Semantic Scene Completion (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.