Research map / 3D vision / 3D generation
Disentangling Instance and Scene Contexts for 3D Semantic Scene Completion
Enyu Liu, En Yu, Sijia Chen, Wenbing Tao
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 instance stream to baseline (ablation) | BEV baseline · SemanticKITTI val | mIoU: 12.69 -> 16.86 | Improved |
| add scene stream to baseline (ablation) | BEV baseline · SemanticKITTI val | mIoU: 12.69 -> 16.73 | Improved |
| combine both streams (ablation) | DISC · SemanticKITTI val | mIoU: 16.86 -> 17.05 | Improved |
| add instance stream to baseline (ablation) | BEV baseline · SemanticKITTI val | InsM: 6.06 -> 9.79 | Improved |
| add coarse-to-fine BEV generation (ablation) | DISC · SemanticKITTI val | mIoU: 16.70 -> 17.05 | Improved |
| add image-assisted candidate localization (ablation) | DISC instance query · SemanticKITTI val | mIoU: 16.64 -> 17.05 | Improved |
| use grid-based neighbor suppression instead of fixed threshold (ablation) | DISC instance query · SemanticKITTI val | mIoU: 16.88 -> 17.05 | Improved |
| add instance-image cross-attention to adaptive instance layer (ablation) | AIL · SemanticKITTI val | mIoU: 16.69 -> 16.98 | Improved |
| add instance-scene cross-attention to AIL (ablation) | AIL · SemanticKITTI val | mIoU: 16.98 -> 17.05 | Improved |
| add instance self-attention alone (ablation) | AIL · SemanticKITTI val | mIoU: 16.73 -> 16.69 | No clear effect |
| add scene self-attention to global scene layer (ablation) | GSL · SemanticKITTI val | mIoU: 16.85 -> 17.05 | Improved |
| add scene-image cross-attention to GSL (ablation) | GSL · SemanticKITTI val | mIoU: 17.01 -> 17.05 | Improved |
Closest papers by meaning
- Semantic Scene Completion via Integrating Instances and Scene In-the-Loop (CVPR 2021 · similarity 0.91)
- Symphonize 3D Semantic Scene Completion with Contextual Instance Queries (CVPR 2024 · similarity 0.90)
- Context and Geometry Aware Voxel Transformer for Semantic Scene Completion (NeurIPS 2024 · similarity 0.90)
- IPFormer: Visual 3D Panoptic Scene Completion with Context-Adaptive Instance Proposals (NeurIPS 2025 · similarity 0.90)
- Memory-Augmented Re-Completion for 3D Semantic Scene Completion (AAAI 2025 · 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.
