Research map / Point clouds / Point cloud classification
A Closer Look at Local Aggregation Operators in Point Cloud Analysis
Ze Liu(†), Han Hu, Yue Cao, Zheng Zhang, Xin Tong
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 1 FC layer instead of 3 FC layers (ablation) | point-wise MLP operator · S3DIS Area-5 | mIoU: 59.9 -> 62.9 | Improved |
| use 1 FC layer instead of 2 FC layers (ablation) | point-wise MLP operator · S3DIS Area-5 | mIoU: 59.5 -> 62.9 | Improved |
| use MAX reduction instead of AVG or SUM (ablation) | point-wise MLP operator · S3DIS Area-5 | mIoU: 44.7 (SUM), 55.1 (AVG) -> 62.9 (MAX) | Improved |
| use 1 FC layer instead of 3 FC layers (ablation) | adaptive weight operator · S3DIS Area-5 | mIoU: 58.5 -> 62.6 | Improved |
| use relative position input instead of edge feature input (ablation) | adaptive weight operator · S3DIS Area-5 | mIoU: 46.6 -> 62.6 | Improved |
| add SoftMax normalization of aggregation weights (ablation) | adaptive weight operator · S3DIS Area-5 | mIoU: 62.6 -> 55.9 | Hurt |
| use AVG reduction instead of SUM (ablation) | adaptive weight operator · S3DIS Area-5 | mIoU: 61.7 -> 62.6 | Improved |
| add neighbor position encoding via PosPool over plain MAX pooling | PosPool* · PartNet test | mIoU: 47.4 -> 53.8 | Improved |
| use PosPool* instead of previous best PointCNN | PosPool* · PartNet test | mIoU: 46.4 -> 53.8 | Improved |
| use full relative position instead of angle or distance alone (ablation) | PosPool variants · PartNet val | mIoU: 43.4 -> 48.5 | Improved |
| use small 16x cheaper configuration (S) instead of regular for PosPool (ablation) | PosPool · S3DIS Area-5 | mIoU: 66.5 -> 64.2 | Hurt |
| use Pseudo grid instead of point-wise MLP at regular size | operator choice · ModelNet40 | accuracy: 92.8 -> 93.0 | No clear effect |
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 3 of 3 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
- DeepLA-Net: Very Deep Local Aggregation Networks for Point Cloud Analysis (CVPR 2025 · similarity 0.89)
- Rethinking Network Design and Local Geometry in Point Cloud: A Simple Residual MLP Framework (ICLR 2022 · similarity 0.88)
- Efficient Point Cloud Processing With High-Dimensional Positional Encoding and Non-Local MLPs (TPAMI 2026 · similarity 0.87)
- Improved MLP Point Cloud Processing with High-Dimensional Positional Encoding (AAAI 2024 · similarity 0.85)
- PLATYPUS: Progressive Local Surface Estimator for Arbitrary-Scale Point Cloud Upsampling (AAAI 2025 · similarity 0.84)
- Point Deformable Network with Enhanced Normal Embedding for Point Cloud Analysis (AAAI 2024 · similarity 0.84)
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.
