Research map / Point clouds / Point cloud classification
Robust 3D Shape Classification via Non-Local Graph Attention Network
Read the paperCheck 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 NLGAT instead of Triangle-Net | 3D shape classification · ModelNet40, 1024 points, arbitrary SO(3) rotation | accuracy: 92.20 vs 86.66 | Helped |
| Use NLGAT instead of Triangle-Net | 3D shape classification · ModelNet40, 64 points, arbitrary SO(3) rotation | accuracy: 89.78 vs 81.53 | Helped |
| Use NLGAT instead of Triangle-Net | 3D shape classification · ModelNet40, 8 points, arbitrary SO(3) rotation | accuracy: 77.20 vs 48.19 | Helped |
| Use NLGAT instead of DGCNN | 3D shape classification · PB T50 RS, 32 points, no SO(3) rotation | accuracy: 77.56 vs 70.70 | Helped |
| Use NLGAT instead of Triangle-Net | 3D shape classification · PB T50 RS, 32 points, SO(3) rotation | accuracy: 76.84 vs 70.16 | Helped |
| Use NLGAT instead of Triangle-Net | 3D shape classification · PB T50 RS, 2048 points, SO(3) rotation | accuracy: 83.30 vs 73.77 | Helped |
| Add GRN to the model (ablation) | NLGAT · ModelNet40, arbitrary orientations | accuracy: 2.5 | Helped |
| Add GRN to the model (ablation) | NLGAT · ModelNet40, z-axis train / SO(3) test | accuracy: 3.2 | Helped |
| Add GSN to the model (ablation) | NLGAT · ModelNet40, arbitrary orientations | accuracy: 3.6 | Helped |
| Add GSN to the model (ablation) | NLGAT · ModelNet40, z-axis train / SO(3) test | accuracy: 4.9 | Helped |
| Use MLP-ST and the GRN rather than removing GRN (ablation) | NLGAT · ModelNet40, 16 points with noise | accuracy: 6.49% | Helped |
| Use the GSN rather than removing GRN and GSN (ablation) | NLGAT · ModelNet40, 8 points with noise | accuracy: 10.48% | Helped |
Closest papers by meaning
- View-GCN: View-Based Graph Convolutional Network for 3D Shape Analysis (CVPR 2020 · similarity 0.88)
- Geometry Sharing Network for 3D Point Cloud Classification and Segmentation (AAAI 2020 · similarity 0.86)
- Point2Sequence: Learning the Shape Representation of 3D Point Clouds with an Attention-Based Sequence to Sequence Network (AAAI 2019 · similarity 0.85)
- GVCNN: Group-View Convolutional Neural Networks for 3D Shape Recognition (CVPR 2018 · similarity 0.85)
- Attentional ShapeContextNet for Point Cloud Recognition (CVPR 2018 · similarity 0.85)
- DGC-GNN: Leveraging Geometry and Color Cues for Visual Descriptor-Free 2D-3D Matching (CVPR 2024 · similarity 0.84)
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.