Research map / Point clouds / Point cloud classification
PointConv: Deep Convolutional Networks on 3D Point Clouds
Check 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 |
|---|---|---|---|
| add inverse density scale (ablation) | PointConv · ScanNet validation, xyz input | mIoU: 60.3 -> 61.0; 58.2 -> 59.0; 56.9 -> 58.2 | Helped |
| apply raw density without the nonlinear MLP transform (ablation) | PointConv · ScanNet validation, xyz input | mIoU: 60.3 -> 60.1; 58.2 -> 57.7; 56.9 -> 57.3 | Hurt |
| use stride 0.5m instead of 1.0m or 1.5m (ablation) | PointConv · ScanNet validation, xyz input | mIoU: 59.0 -> 61.0; 58.2 -> 61.0 | Helped |
| add RGB input features (ablation) | PointConv · ScanNet validation, stride 0.5m | mIoU: 61.0 -> 60.8 | No clear effect |
| add RGB input features (ablation) | PointConv · ScanNet validation, stride 1.0m | mIoU: 59.0 -> 58.6 | Hurt |
| add RGB input features (ablation) | PointConv · ScanNet validation, stride 1.5m | mIoU: 58.2 -> 57.5 | Hurt |
| increase the MLP intermediate width or vary its layer count (ablation) | PointConv · ScanNet subset classification | accuracy | No clear effect |
Closest papers by meaning
- PointConvFormer: Revenge of the Point-Based Convolution (CVPR 2023 · similarity 0.93)
- PointCNN++: Performant Convolution on Native Points (CVPR 2026 · similarity 0.92)
- PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation (CVPR 2017 · similarity 0.91)
- Pointwise Convolutional Neural Networks (CVPR 2018 · similarity 0.91)
- LinNet: Linear Network for Efficient Point Cloud Representation Learning (NeurIPS 2024 · similarity 0.88)
- DeepLA-Net: Very Deep Local Aggregation Networks for Point Cloud Analysis (CVPR 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.