Research map / Point clouds / Point cloud classification
PRIN/SPRIN: On Extracting Point-Wise Rotation Invariant Features
Yang You, Yujing Lou, Ruoxi Shi, Qi Liu, Yu‐Wing Tai, Lizhuang Ma, Weiming Wang, Cewu Lu
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 |
|---|---|---|---|
| use density aware adaptive sampling (DAAS) instead of uniform sampling (ablation) | PRIN · ShapeNet part, rotated | mIoU: 67.47 -> 71.20 | Improved |
| use density aware adaptive sampling (DAAS) instead of uniform sampling (ablation) | PRIN · ShapeNet part, rotated | accuracy: 83.13 -> 85.99 | Improved |
| increase bandwidth from 4 to 32 (ablation) | PRIN · ShapeNet part, rotated | accuracy and mIoU: 72.20/52.98 -> 85.99/71.20 | Improved |
| increase bandwidth from 16 to 32 (ablation) | PRIN · ShapeNet part, rotated | mIoU: 68.09 -> 71.20 | Improved |
| add triangle features to correlation features (ablation) | SPRIN · ShapeNet part, rotated | mIoU: 79.70 -> 82.67 | Improved |
| add correlation features to triangle features (ablation) | SPRIN · ShapeNet part, rotated | mIoU: 81.38 -> 82.67 | Improved |
| reduce input points from 2048 to 1024 (ablation) | SPRIN · ShapeNet part, rotated | mIoU: 82.67 -> 79.79 | Hurt |
| reduce input points from 2048 to 256 (ablation) | PRIN · ShapeNet part, rotated | mIoU: 71.20 -> 36.80 | Hurt |
| use sparse SPRIN instead of dense PRIN | PRIN · ShapeNet part, rotated | avg instance IoU: 71.20 -> 82.67 | Improved |
| test on rotated rather than non-rotated shapes | SPRIN · ShapeNet part | avg instance IoU: 82.59 -> 82.67 | No clear effect |
| test on rotated rather than non-rotated shapes | PointNet++ · ShapeNet part | avg instance IoU: 84.63 -> 36.66 | Hurt |
| test on arbitrary rotations rather than none | SPRIN · ModelNet40 | accuracy: 86.01 -> 86.13 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 2 on most shared results.
- SO-Net — SO-Net: Self-Organizing Network for Point Cloud Analysis (CVPR 2018) · better on 0 of 1 shared result
- SpiderCNN — SpiderCNN: Deep Learning on Point Sets with Parameterized Convolutional Filters (ECCV 2018) · better on 0 of 1 shared result
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
- Pointwise Rotation-Invariant Network with Adaptive Sampling and 3D Spherical Voxel Convolution (AAAI 2020 · similarity 0.95)
- RIConv++: Effective Rotation Invariant Convolutions for 3D Point Clouds Deep Learning (IJCV 2022 · similarity 0.89)
- RISurConv: Rotation Invariant Surface Attention-Augmented Convolutions for 3D Point Cloud Classification and Segmentation (ECCV 2024 · similarity 0.89)
- Hierarchical Direction Perception via Atomic Dot-Product Operators for Rotation-Invariant Point Clouds Learning (AAAI 2026 · similarity 0.88)
- Deep Positional and Relational Feature Learning for Rotation-Invariant Point Cloud Analysis (ECCV 2020 · similarity 0.88)
- CRIN: Rotation-Invariant Point Cloud Analysis and Rotation Estimation via Centrifugal Reference Frame (AAAI 2023 · similarity 0.88)
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.
