Research map / Point clouds / Segmentation and detection
Geometric-Aware Hypergraph Reasoning for Novel Class Discovery in Point Cloud Segmentation
Zihao Zhang, Aming Wu, Yang Li, Yahong Han, Jialie Shen
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 Geometry-Aware Prototype to baseline (ablation) | MinkowskiUNet-34C baseline · SemanticPOSS split 0 overall Avg | mIoU: 23.2 -> 27.4 | Improved |
| add Hypergraph Structure Construction to baseline (ablation) | MinkowskiUNet-34C baseline · SemanticPOSS split 0 overall Avg | mIoU: 23.2 -> 32.8 | Improved |
| add GAP on top of hypergraph (geometric-aware instead of clustered prototypes) (ablation) | HSC model · SemanticPOSS split 0 overall Avg | mIoU: 32.8 -> 34.2 | Improved |
| add Dynamic Hyperedge Adjustment Mechanism (ablation) | GAP + HSC · SemanticPOSS split 0 overall Avg | mIoU: 34.2 -> 36.5 | Improved |
| add Dynamic Hyperedge Adjustment Mechanism (ablation) | GAP + HSC · SemanticPOSS split 0 novel Avg | mIoU: 45.2 -> 47.2 | Improved |
| add DHAM on car class (ablation) | GAP + HSC · SemanticPOSS split 0 car | IoU: 6.5 -> 4.4 | Hurt |
| use hypergraph reasoning instead of DASL | Ours · SemanticPOSS split 2 novel | mIoU: 12.6 -> 22.3 | Improved |
| use hypergraph reasoning instead of DASL | Ours · SemanticPOSS split 3 novel | mIoU: 17.7 -> 37.8 | Improved |
| use hypergraph reasoning instead of DASL | Ours · SemanticPOSS split 0 novel | mIoU: 48.4 -> 47.2 | Mixed |
| use hypergraph reasoning instead of DASL | Ours · SemanticKITTI split 3 novel | mIoU: 12.6 -> 11.8 | Hurt |
| use hypergraph reasoning instead of DASL | Ours · SemanticKITTI split 2 novel | mIoU: 20.1 -> 20.9 | No clear effect |
| increase neighboring prototypes M from 8 to 10 (ablation) | Ours · SemanticPOSS and SemanticKITTI | mIoU | Hurt |
Closest papers by meaning
- Geometry and Uncertainty-Aware 3D Point Cloud Class-Incremental Semantic Segmentation (CVPR 2023 · similarity 0.91)
- Novel Class Discovery for Point Cloud Segmentation via Joint Learning of Causal Representation and Reasoning (NeurIPS 2025 · similarity 0.90)
- PDF: A Probability-Driven Framework for Open World 3D Point Cloud Semantic Segmentation (CVPR 2024 · similarity 0.87)
- Geometrically-driven Aggregation for Zero-shot 3D Point Cloud Understanding (CVPR 2024 · similarity 0.87)
- LogoSP: Local-global Grouping of Superpoints for Unsupervised Semantic Segmentation of 3D Point Clouds (CVPR 2025 · similarity 0.87)
- Hyperbolic Uncertainty-Aware Few-Shot Incremental Point Cloud Segmentation (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.
