Research map / Point clouds / Point cloud classification
Zero-Shot Learning on 3D Point Cloud Objects and Beyond
Ali Cheraghian, Shafin Rahman, Townim Faisal Chowdhury, Dylan Campbell, Lars Petersson
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 |
|---|---|---|---|
| project semantics to point-cloud features (S2F) instead of F2S (ablation) | inductive PointConv ZSL model · ModelNet10 ZSL | accuracy: 24.45 -> 32.49 | Improved |
| add unsupervised triplet loss (transductive) (ablation) | S2F model · ModelNet10 ZSL | accuracy: 32.49 -> 43.17 | Improved |
| add hubness loss (ablation) | transductive model · ModelNet10 ZSL | accuracy: 43.17 -> 61.45 | Improved |
| add unbiasing loss (ablation) | transductive model · ModelNet10 GZSL | harmonic mean: 57.33 -> 73.39 | Improved |
| add unbiasing loss (ablation) | transductive model · ModelNet10 ZSL | accuracy: 61.45 -> 68.50 | Improved |
| add unbiasing loss (ablation) | transductive model · ScanObjectNN ZSL | accuracy: 30.74 -> 30.53 | No clear effect |
| use confidence weighting pi in hubness loss instead of pi=1 (ablation) | transductive model · ModelNet10 ZSL | accuracy: 46.37 -> 68.50 | Improved |
| use confidence weighting pi in hubness loss instead of pi=1 (ablation) | transductive model · ScanObjectNN GZSL | harmonic mean: 28.19 -> 45.63 | Improved |
| add transductive training with unlabeled unseen data | Ours · ModelNet10 GZSL | harmonic mean: 12.69 -> 73.39 | Improved |
| use PointConv backbone instead of PointNet | Ours (Transductive) · ModelNet10 ZSL | accuracy: 18.28 -> 68.50 | Improved |
| use PointConv backbone instead of DGCNN | Ours (Transductive) · ModelNet10 ZSL | accuracy: 60.05 -> 68.50 | Improved |
| add transductive learning with few seen classes (20 vs 30) | Ours (Transductive) · ModelNet10 ZSL | accuracy: 68.50 -> 27.42 | Hurt |
Closest papers by meaning
- A Closer Look at Few-Shot 3D Point Cloud Classification (IJCV 2022 · similarity 0.92)
- Generalized Zero-Shot Learning for Point Cloud Segmentation with Evidence-Based Dynamic Calibration (AAAI 2025 · similarity 0.90)
- Point2Real: Bridging the Gap between Point Cloud and Realistic Image for Open-World 3D Recognition (AAAI 2024 · similarity 0.88)
- Geometrically-driven Aggregation for Zero-shot 3D Point Cloud Understanding (CVPR 2024 · similarity 0.88)
- Generalized Few-shot 3D Point Cloud Segmentation with Vision-Language Model (CVPR 2025 · similarity 0.87)
- Data Distribution Distilled Generative Model for Generalized Zero-Shot Recognition (AAAI 2024 · 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.
