Research map / Point clouds / Point cloud classification
A Unified Query-Based Paradigm for Point Cloud Understanding
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 |
|---|---|---|---|
| Replace the Sparse U-Net decoder with hierarchical Q-Net querying | Sparse U-Net · ScanNet validation | mIoU (%): 72.9 -> 75.3; +2.4 | Helped |
| Replace the Sparse U-Net decoder with hierarchical Q-Net querying | Sparse U-Net · S3DIS Area 5 | mIoU (%): 66.9 -> 71.3; +4.4 | Helped |
| Replace the Sparse U-Net decoder with hierarchical Q-Net querying | Sparse U-Net · S3DIS 6-fold | mIoU (%): 72.6 -> 77.5; +4.9 | Helped |
| Replace PointNet++ with EQ-PointNet++ | VoteNet · ScanNetV2 | mAP@0.5: 39.9 -> 45.4; 5.5 % | Helped |
| Replace PointNet++ with EQ-PointNet++ | GroupFree · ScanNetV2 | mAP@0.25: 67.3 -> 68.0; 0.7 % | Helped |
| Replace PointNet++ with EQ-PointNet++ | VoteNet · SUN RGB-D | mAP@0.5: 32.9 -> 38.5; 2.7 % | Helped |
| Apply EQ-Paradigm to PointNet++ | PointNet++ (SSG) · ModelNet40 | Accuracy (%): 92.1 -> 93.2; 1.1 % | Helped |
| Add hierarchical feature levels to Q-Net (ablation) | Sparse EQ-Net · ScanNet validation | mIoU (%): 58.4 -> 64.1 -> 68.3 -> 71.9 -> 74.2 -> 75.3 | Helped |
| Add query self-attention to the Q-Decoder (ablation) | EQ-SECOND · KITTI val, Car Moderate, random query test | AP (%): 81.49 -> 74.96; 6.65 % | Hurt |
| Add query self-attention to the Q-Decoder (ablation) | EQ-SECOND · KITTI val, Car Moderate, patch query test | AP (%): 81.49 -> 81.61; 0.12 % | Helped |
| Use a point-based embedding network instead of a voxel-based embedding network (ablation) | SECOND head under EQ-Paradigm · KITTI val, Car Moderate | AP (%): 81.49 -> 82.70; 1.21 % | Helped |
| Combine voxel-based and point-based embedding networks (ablation) | SECOND head under EQ-Paradigm · KITTI val, Car Moderate | AP (%): 82.70 -> 82.94 | Helped |
Closest papers by meaning
- Interpretable3D: An Ad-Hoc Interpretable Classifier for 3D Point Clouds (AAAI 2024 · similarity 0.86)
- Geometrically-driven Aggregation for Zero-shot 3D Point Cloud Understanding (CVPR 2024 · similarity 0.85)
- Sonata: Self-Supervised Learning of Reliable Point Representations (CVPR 2025 · similarity 0.85)
- PMA: Towards Parameter-Efficient Point Cloud Understanding via Point Mamba Adapter (CVPR 2025 · similarity 0.85)
- Exploring Vision Semantic Prompt for Efficient Point Cloud Understanding (ICML 2025 · similarity 0.84)
- DOCTR: Disentangled Object-Centric Transformer for Point Scene Understanding (AAAI 2024 · similarity 0.84)
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.