Research map / Point clouds / Segmentation and detection
SCF-Net: Learning Spatial Contextual Features for Large-Scale 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 |
|---|---|---|---|
| Use SCF-Net instead of the strongest listed baseline | Point-cloud semantic segmentation · S3DIS 6-fold cross-validation | mIoU: 70.6 -> 71.6 | Helped |
| Use SCF-Net instead of the strongest listed baseline | Point-cloud semantic segmentation · S3DIS 6-fold cross-validation | mAcc: 82.0 -> 82.7 | Helped |
| Use SCF-Net instead of the strongest listed baseline | Point-cloud semantic segmentation · S3DIS 6-fold cross-validation | OA: 88.1 -> 88.4 | Helped |
| Use SCF-Net instead of RandLA-Net | Point-cloud semantic segmentation · Semantic3D reduced-8 test split | mIoU: 77.4 -> 77.6 | Helped |
| Use SCF-Net instead of RandLA-Net | Point-cloud semantic segmentation · Semantic3D reduced-8 test split | OA: 94.8 -> 94.7 | Hurt |
| Add the global contextual feature block (GCF) (ablation) | SCF-Net · S3DIS 6-fold cross-validation | mIoU: 70.5 -> 71.6 | Helped |
| Add dual-distance attentive pooling (DDAP) instead of SAP (ablation) | SCF-Net without GCF · S3DIS 6-fold cross-validation | mIoU: 69.3 -> 70.5 | Helped |
| Add local polar representation (LPR) instead of LSR (ablation) | SCF-Net without GCF and with SAP · S3DIS 6-fold cross-validation | mIoU: 67.9 -> 69.3 | Helped |
| Retain the feature distance in DDAP (ablation) | DDAP · S3DIS area 2 | mIoU: 59.3 -> 59.7 | Helped |
| Add geometric distance to DDAP (ablation) | DDAP without geometric and feature distances · S3DIS area 2 | mIoU: 57.8 -> 59.7 | Helped |
| Concatenate dual-distance and neighbor features instead of weighted summation (ablation) | DDAP · S3DIS area 2 | mIoU: 58.9 -> 59.7 | Helped |
| Set λ to 0.1 rather than 1 (ablation) | SCF-Net · S3DIS areas 1–6 | mIoU: 74.4 -> 75.1; 59.4 -> 59.7; 76.7 -> 78.4; 58.0 -> 60.2; 62.0 -> 63.4; 79.0 -> 80.2 | Mixed |
Closest papers by meaning
- Exploiting Local and Global Structure for Point Cloud Semantic Segmentation with Contextual Point Representations (NeurIPS 2019 · similarity 0.88)
- SSPC-Net: Semi-supervised Semantic 3D Point Cloud Segmentation Network (AAAI 2021 · similarity 0.88)
- Large-Scale Point Cloud Semantic Segmentation With Superpoint Graphs (CVPR 2018 · similarity 0.88)
- Improving Graph Representation for Point Cloud Segmentation via Attentive Filtering (CVPR 2023 · similarity 0.87)
- Taylor Series-Inspired Local Structure Fitting Network for Few-shot Point Cloud Semantic Segmentation (AAAI 2025 · similarity 0.85)
- PDF: A Probability-Driven Framework for Open World 3D Point Cloud Semantic Segmentation (CVPR 2024 · similarity 0.85)
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.