Research map / Point clouds / Point cloud classification
Attentional ShapeContextNet for Point Cloud Recognition
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 |
|---|---|---|---|
| Replace independent point features with hierarchical shape-context blocks | PointNet-style point cloud classification on MNIST · MNIST test (256 points) | error rate (%): 0.78 -> 0.60 | Helped |
| Use a five-block ShapeContextNet instead of a single shape-context block (ablation) | ShapeContextNet · MNIST test (256 points) | error rate (%) | Helped |
| Replace hand-designed shape-context blocks with self-attention blocks | Point cloud classification · ModelNet40 test | overall accuracy (%): 90.0 -> 89.8 | No clear effect |
| Use 2 attention heads instead of 1 (ablation) | A-SCN · ModelNet40 test | overall accuracy (%): 89.8 -> 89.2 | Hurt |
| Use 4 attention heads instead of 1 (ablation) | A-SCN · ModelNet40 test | overall accuracy (%): 89.8 -> 89.8 | No clear effect |
| Use hierarchical contextual modeling for semantic segmentation | PointNet · S3DIS | mean IoU (%): 47.71 -> 52.72 | Helped |
| Use hierarchical contextual modeling for part segmentation | PointNet · ShapeNet part test | mean IoU (%): 83.7 -> 84.6 | Helped |
Closest papers by meaning
- Shape-Oriented Convolution Neural Network for Point Cloud Analysis (AAAI 2020 · similarity 0.90)
- Point2Sequence: Learning the Shape Representation of 3D Point Clouds with an Attention-Based Sequence to Sequence Network (AAAI 2019 · similarity 0.90)
- PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation (CVPR 2017 · similarity 0.89)
- PCAN: 3D Attention Map Learning Using Contextual Information for Point Cloud Based Retrieval (CVPR 2019 · similarity 0.89)
- PointAttN: You Only Need Attention for Point Cloud Completion (AAAI 2024 · similarity 0.87)
- PointCFormer: A Relation-Based Progressive Feature Extraction Network for Point Cloud Completion (AAAI 2025 · similarity 0.87)
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.