Research map / Point clouds / Segmentation and detection
Uncertainty-Driven Gaussian Sphere Propagation for 3D Semantic Segmentation
Zhicheng Yan, Qingyong Li, Yixiao Song, Wen Wang
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 UGSP uncertainty-driven Gaussian sphere propagation | PTv3 · S3DIS Area 5 | mIoU: 73.3 -> 75.6 | Improved |
| add UGSP | PTv3 · ScanNet V2 | mIoU: 76.8 -> 77.9 | Improved |
| add UGSP | PTv3 · SemanticKITTI | mIoU: 72.3 -> 73.8 | Improved |
| add UGSP | PTv3 · Waymo | mIoU: 71.3 -> 72.1 | Mixed |
| use full uncertainty-driven region growing instead of RAND-KNN (ablation) | UGSP · S3DIS Area 5 | mIoU: 72.8 -> 75.6 | Improved |
| use full uncertainty region growth instead of uncertainty-only initialization (ablation) | UGSP · S3DIS Area 5 | mIoU: 74.1 -> 75.6 | Improved |
| use SH encoding instead of max pooling (ablation) | UGSP · S3DIS Area 5 | mIoU: 72.8 -> 75.6 | Improved |
| use SH encoding instead of average pooling (ablation) | UGSP · S3DIS Area 5 | mIoU: 73.1 -> 75.6 | Improved |
| add attention with path-direction modeling (ablation) | UGSP · S3DIS Area 5 | mIoU: 73.3 -> 75.6 | Improved |
| use attention instead of concatenation (ablation) | UGSP · S3DIS Area 5 | mIoU: 73.5 -> 75.6 | Improved |
| raise SH order from 2 to 3 (ablation) | UGSP · S3DIS Area 5 | mIoU: 74.23 -> 75.62 | Improved |
| raise SH order from 3 to 4 (ablation) | UGSP · S3DIS Area 5 | mIoU: 75.62 -> 75.64 | No clear effect |
Closest papers by meaning
- AdapToPASS: Ambiguity-aware Adaptive Spherical Transformer for Panoramic Semantic Segmentation (NeurIPS 2026 · similarity 0.87)
- Bootstraping Clustering of Gaussians for View-consistent 3D Scene Understanding (AAAI 2025 · similarity 0.87)
- Geometrically-driven Aggregation for Zero-shot 3D Point Cloud Understanding (CVPR 2024 · similarity 0.87)
- Generalized Zero-Shot Learning for Point Cloud Segmentation with Evidence-Based Dynamic Calibration (AAAI 2025 · similarity 0.87)
- PDF: A Probability-Driven Framework for Open World 3D Point Cloud Semantic Segmentation (CVPR 2024 · similarity 0.86)
- LogoSP: Local-global Grouping of Superpoints for Unsupervised Semantic Segmentation of 3D Point Clouds (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.
