Research map / Point clouds / Point cloud classification
Point2Sequence: Learning the Shape Representation of 3D Point Clouds with an Attention-Based Sequence to Sequence 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 attention with the encoder-decoder rather than removing attention (ablation) | Point2Sequence · ModelNet40 | instance accuracy: 92.26 -> 92.54 | Helped |
| include the decoder rather than removing it (ablation) | Point2Sequence · ModelNet40 | instance accuracy: 92.42 -> 92.54 | Helped |
| use the encoder-decoder rather than concatenation to aggregate area features (ablation) | Point2Sequence · ModelNet40 | instance accuracy: 92.06 -> 92.54 | Helped |
| use the encoder-decoder rather than max pooling to aggregate area features (ablation) | Point2Sequence · ModelNet40 | instance accuracy: 91.73 -> 92.54 | Helped |
| use multiple area scales rather than a single scale (ablation) | Point2Sequence · ModelNet40 | instance accuracy: 91.94 -> 92.54 | Helped |
| use 2 area scales rather than 4 (ablation) | Point2Sequence · ModelNet40 | instance accuracy: 92.54 -> 92.63 | Helped |
| use 384 sampled centroids rather than 128 or 512 (ablation) | Point2Sequence · ModelNet40 | instance accuracy: 91.86 -> 92.54 | Helped |
| use an LSTM cell rather than a GRU cell (ablation) | Point2Sequence · ModelNet40 | instance accuracy: 92.18 -> 92.54 | Helped |
| use hidden-state dimension 128 rather than 64 or 256 (ablation) | Point2Sequence · ModelNet40 | instance accuracy: 92.46 -> 92.54; 92.18 -> 92.54 | Helped |
| use learning rate 0.001 rather than 0.0005 or 0.002 (ablation) | Point2Sequence · ModelNet40 | instance accuracy: 91.94 -> 92.63; 91.86 -> 92.63 | Helped |
Closest papers by meaning
- Attentional ShapeContextNet for Point Cloud Recognition (CVPR 2018 · similarity 0.90)
- PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation (CVPR 2017 · similarity 0.89)
- Shape-Oriented Convolution Neural Network for Point Cloud Analysis (AAAI 2020 · similarity 0.88)
- PointAttN: You Only Need Attention for Point Cloud Completion (AAAI 2024 · similarity 0.87)
- Multi-View Representation is What You Need for Point-Cloud Pre-Training (ICLR 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.