Research map / Point clouds / Segmentation and detection
SCPNet: Semantic Scene Completion on Point Cloud
Read the paperCheck 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 4 input frames instead of a single frame | SCPNet · SemanticKITTI test | mIoU: 36.7 -> 47.5 | Helped |
| Add DSKD relation-based distillation (ablation) | SCPNet · SemanticKITTI val | mIoU: 34.4 -> 37.2 | Helped |
| Replace DSKD with vanilla KD (ablation) | SCPNet · SemanticKITTI val | mIoU: 34.4 -> 33.8 | Hurt |
| Replace DSKD with FitNets (ablation) | SCPNet · SemanticKITTI val | mIoU: 34.4 -> 33.8 | Hurt |
| Replace DSKD with PKT (ablation) | SCPNet · SemanticKITTI val | mIoU: 34.4 -> 34.6 | Helped |
| Replace DSKD with PVKD (ablation) | SCPNet · SemanticKITTI val | mIoU: 34.4 -> 36.2 | Helped |
| Replace DSKD with NST (ablation) | SCPNet · SemanticKITTI val | mIoU: 34.4 -> 36.3 | Helped |
| Add the completion sub-network (ablation) | JS3C-Net · SemanticKITTI val | mIoU: 24.0 -> 30.8 | Helped |
| Add downsampling to the completion sub-network (ablation) | SCPNet · SemanticKITTI val | mIoU: 37.2 -> 33.1 | Hurt |
| Rectify completion labels using panoptic instance labels (ablation) | SCPNet · SemanticKITTI val | mIoU: 37.2 -> 40.8 | Helped |
| Set DSKD loss coefficient β from 3000 to 4000 (ablation) | SCPNet · SemanticKITTI val | mIoU: 37.2 -> 35.1 | Hurt |
| Set DSKD loss coefficient β from 3000 to 2000 (ablation) | SCPNet · SemanticKITTI val | mIoU: 37.2 -> 35.3 | Hurt |
Closest papers by meaning
- Voxel Proposal Network via Multi-Frame Knowledge Distillation for Semantic Scene Completion (NeurIPS 2024 · similarity 0.89)
- Memory-Augmented Re-Completion for 3D Semantic Scene Completion (AAAI 2025 · similarity 0.87)
- VLScene: Vision-Language Guidance Distillation for Camera-Based 3D Semantic Scene Completion (AAAI 2025 · similarity 0.87)
- Context and Geometry Aware Voxel Transformer for Semantic Scene Completion (NeurIPS 2024 · similarity 0.87)
- Semantic Complete Scene Forecasting from a 4D Dynamic Point Cloud Sequence (AAAI 2024 · similarity 0.87)
- Semi-supervised 3D Semantic Scene Completion with 2D Vision Foundation Model Guidance (AAAI 2025 · similarity 0.87)
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.