Research map / Point clouds / Point cloud classification
3D Point Capsule Networks
Yongheng Zhao, Tolga Birdal, Haowen Deng, Federico Tombari
Check 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 |
|---|---|---|---|
| Use 3D-PointCapsNet instead of PPF-FoldNet | Unsupervised local feature extraction · 3DMatch | recall (Average): 0.7182 -> 0.7867 | Improved |
| Use 3D-PointCapsNet instead of PPF-FoldNet | Unsupervised local feature extraction · Rotated-3DMatch | recall (Average): 0.7311 -> 0.8074 | Improved |
| Use multiple latent capsules rather than a single latent vector | 3D-PointCapsNet · ShapeNet-Core v2 | Chamfer distance (×10^3): 1.51 -> 1.46 | Improved |
| Use latent capsules learned on ShapeNet-Parts for transfer | 3D object classification · ModelNet40 | accuracy: 88.4 -> 89.3 | Improved |
| Use capsule-wise part prediction | Part segmentation · ShapeNet-Part (1% training data) | accuracy: 0.78 -> 0.85 | Improved |
| Use capsule-wise part prediction | Part segmentation · ShapeNet-Part (5% training data) | IoU: 0.69 -> 0.70 | Improved |
| Replace dynamic routing with PointNet-like convolutional layers (ablation) | 3D-PointCapsNet auto-encoder · 3D shape reconstruction | reconstruction quality | Hurt |
| Use capsule-based segmentation rather than SO-Net | Part segmentation · ShapeNet-Part (5% training data) | accuracy: 0.84 -> 0.86 | Improved |
| Use capsule-based segmentation rather than FoldingNet | Part segmentation · ShapeNet-Part (1% training data) | accuracy: 56.15 -> 59.24 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- FoldingNet — FoldingNet: Point Cloud Auto-Encoder via Deep Grid Deformation (CVPR 2018) · better on 6 of 6 shared results
- SONet — SO-Net: Self-Organizing Network for Point Cloud Analysis (CVPR 2018) · better on 4 of 4 shared results
- PPF-FoldNet-5K — PPF-FoldNet: Unsupervised Learning of Rotation Invariant 3D Local Descriptors (ECCV 2018) · better on 2 of 2 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
- 3DPointCaps++: Learning 3D Representations with Capsule Networks (IJCV 2022 · similarity 0.89)
- Canonical Capsules: Self-Supervised Capsules in Canonical Pose (NeurIPS 2021 · similarity 0.88)
- Quaternion Equivariant Capsule Networks for 3D Point Clouds (ECCV 2020 · similarity 0.88)
- Fully-Convolutional Point Networks for Large-Scale Point Clouds (ECCV 2018 · similarity 0.87)
- PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation (CVPR 2017 · similarity 0.87)
- CAPNet: Continuous Approximation Projection for 3D Point Cloud Reconstruction Using 2D Supervision (AAAI 2019 · 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.
