Research map / Point clouds / Point cloud classification
ShellNet: Efficient Point Cloud Convolutional Neural Networks Using Concentric Shells Statistics
Zhiyuan Zhang, Binh-Son Hua, Sai-Kit Yeung
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 |
|---|---|---|---|
| increase shell size from 8 to 16 (ablation) | ShellNet · ModelNet40 | overall accuracy (%): 91.0 -> 93.1 | Improved |
| increase shell size from 16 to 32 (ablation) | ShellNet · ModelNet40 | overall accuracy (%): 93.1 -> 93.1 | No clear effect |
| increase shell size from 32 to 64 (ablation) | ShellNet · ModelNet40 | overall accuracy (%): 93.1 -> 92.8 | Hurt |
| use farthest point sampling instead of random sampling (ablation) | ShellNet classification · ModelNet40 | overall accuracy (%): 93.1 -> 93.1 | No clear effect |
| use equidistant shells with dynamic point counts instead of fixed shell size (ablation) | ShellNet classification · ModelNet40 | overall accuracy (%): 93.1 -> 92.7 | Hurt |
| search neighbors in feature space instead of xyz space (ablation) | ShellNet classification · ModelNet40 | overall accuracy (%): 93.1 -> 92.4 | Hurt |
| use farthest point sampling for neighbors in segmentation (ablation) | ShellNet segmentation · part and semantic segmentation | accuracy | Improved |
| reduce shell size to get small receptive field (ablation) | ShellNet · ModelNet40 efficiency | training time per batch: 0.066s -> 0.025s | Improved |
| sparse input point clouds instead of original | ShellNet · ShapeNet part segmentation | mpIoU: 82.4 -> 80.2 | Hurt |
| partial input point clouds instead of original | ShellNet · ShapeNet part segmentation | mpIoU: 82.4 -> 72.6 | Hurt |
Later papers that compared with this method
1 later accepted paper reports ShellNet as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Salient Object Detection for Point Clouds (ECCV 2022 · better on 4 of 4 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- PointCNN — PointCNN: Convolution On X-Transformed Points (NeurIPS 2018) · better on 5 of 6 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- RIConv++: Effective Rotation Invariant Convolutions for 3D Point Clouds Deep Learning (IJCV 2022 · similarity 0.90)
- PointConv: Deep Convolutional Networks on 3D Point Clouds (CVPR 2019 · similarity 0.89)
- PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation (CVPR 2017 · similarity 0.89)
- Pointwise Convolutional Neural Networks (CVPR 2018 · similarity 0.89)
- A-CNN: Annularly Convolutional Neural Networks on Point Clouds (CVPR 2019 · similarity 0.89)
- LinNet: Linear Network for Efficient Point Cloud Representation Learning (NeurIPS 2024 · 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.
