Research map / Point clouds / Segmentation and detection
Pyramid Architecture for Multi-Scale Processing in Point Cloud Segmentation
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 |
|---|---|---|---|
| Add lateral information flow (ablation) | BaseNet · S3DIS Area-5 | mIoU: 66.0 -> 66.6 | Helped |
| Add downward information flow to lateral links (ablation) | BaseNet · S3DIS Area-5 | mIoU: 66.6 -> 67.6 | Helped |
| Add upward information flow to lateral links (ablation) | BaseNet · S3DIS Area-5 | mIoU: 66.6 -> 67.2 | Helped |
| Add both upward and downward information flows to lateral links (ablation) | BaseNet · S3DIS Area-5 | mIoU: 66.0 -> 68.2 | Helped |
| Use CLAIM instead of concatenation for cross-scale fusion (ablation) | Bidirectional pyramid · S3DIS Area-5 | mIoU: 67.9 -> 68.2 | Helped |
| Use identity pre-fusion transforms rather than MLP pre-fusion transforms with concatenation (ablation) | Bidirectional pyramid · S3DIS Area-5 | mIoU: 64.7 -> 67.9 | Helped |
| Use local aggregation rather than MLP as the post-fusion transform (ablation) | Bidirectional pyramid with concatenation fusion · S3DIS Area-5 | mIoU: 67.9 -> 66.6 | Hurt |
| Replace CLAIM with stacked SE attention (ablation) | Bidirectional pyramid · S3DIS Area-5 | mIoU: 68.2 -> 66.2 | Hurt |
| Replace CLAIM with stacked CBAM attention (ablation) | Bidirectional pyramid · S3DIS Area-5 | mIoU: 68.2 -> 65.2 | Hurt |
| Add pyramid architecture to KPConv deform | KPConv deform · NPM3D (PL3D) | mIoU: 75.9 -> 83.0 | Helped |
| Add pyramid architecture to KPConv deform | KPConv deform · S3DIS | mIoU: 70.6 -> 73.0 | Helped |
| Add pyramid architecture to KPConv deform | KPConv deform · Semantic3D | mIoU: 73.1 -> 75.8 | Helped |
Closest papers by meaning
- Point-MoE: Large-Scale Multi-Dataset Training with Mixture-of-Experts for 3D Semantic Segmentation (ICLR 2026 · similarity 0.88)
- Semantic Segmentation for Real Point Cloud Scenes via Bilateral Augmentation and Adaptive Fusion (CVPR 2021 · similarity 0.88)
- RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds (CVPR 2020 · similarity 0.88)
- Towards Semantic Segmentation of Urban-Scale 3D Point Clouds: A Dataset, Benchmarks and Challenges (CVPR 2021 · similarity 0.88)
- Improved MLP Point Cloud Processing with High-Dimensional Positional Encoding (AAAI 2024 · similarity 0.87)
- DeepLA-Net: Very Deep Local Aggregation Networks for Point Cloud Analysis (CVPR 2025 · similarity 0.86)
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.