Research map / Point clouds / Point cloud classification
AGConv: Adaptive Graph Convolution on 3D Point Clouds
Mingqiang Wei, Zeyong Wei, Haoran Zhou, Fei Hu, Huajian Si, Zhilei Chen, Zhe Zhu, Jingbo Qiu, Xuefeng Yan, Yanwen Guo, Jun Wang, Jing Qin
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 adaptive kernels (AGConv) instead of standard graph convolution (ablation) | AGConv network · ShapeNetPart | mIoU (%): 85.5 -> 86.4 | Improved |
| use AGConv instead of point-wise attention (ablation) | AGConv network · ShapeNetPart | mIoU (%): 83.3 -> 86.4 | Improved |
| use AGConv instead of channel-wise attention (ablation) | AGConv network · ShapeNetPart | mIoU (%): 83.0 -> 86.4 | Improved |
| use AGConv instead of standard graph convolution (ablation) | AGConv network · ModelNet40 | OA (%): 92.5 -> 93.4 | Improved |
| use AGConv instead of point attention (ablation) | AGConv network · ModelNet40 | OA (%): 92.1 -> 93.4 | Improved |
| use AGConv instead of channel attention (ablation) | AGConv network · ModelNet40 | OA (%): 92.2 -> 93.4 | Improved |
| convolve kernel with xyz instead of features (ablation) | AGConv network · ShapeNetPart | mIoU (%): 85.9 -> 86.4 | Improved |
| use xyz only instead of normals only (ablation) | AGConv network · ShapeNetPart | mIoU (%): 86.2 -> 86.4 | Improved |
| use xyz only instead of xyz plus normals (ablation) | AGConv network · ShapeNetPart | mIoU (%): 86.1 -> 86.4 | No clear effect |
| add spatial transformer network (ablation) | AGConv segmentation network · ShapeNetPart | mIoU (%): 86.2 -> 86.4 | Improved |
| increase neighbors from k=5 to k=20 (ablation) | AGConv classification network · ModelNet40 | OA (%): 92.8 -> 93.4 | Improved |
| increase neighbors from k=20 to k=40 (ablation) | AGConv classification network · ModelNet40 | OA (%): 93.4 -> 93.0 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 4 on most shared results.
- PointASNL — PointASNL: Robust Point Clouds Processing Using Nonlocal Neural Networks With Adaptive Sampling (CVPR 2020) · better on 4 of 4 shared results
- PointCNN — PointCNN: Convolution On X-Transformed Points (NeurIPS 2018) · better on 2 of 2 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 1 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
- Adaptive Graph Convolution for Point Cloud Analysis (ICCV 2021 · similarity 0.94)
- Learning of 3D Graph Convolution Networks for Point Cloud Analysis (TPAMI 2021 · similarity 0.90)
- Convolution in the Cloud: Learning Deformable Kernels in 3D Graph Convolution Networks for Point Cloud Analysis (CVPR 2020 · similarity 0.90)
- PointConv: Deep Convolutional Networks on 3D Point Clouds (CVPR 2019 · similarity 0.90)
- diffConv: Analyzing Irregular Point Clouds with an Irregular View (ECCV 2022 · similarity 0.90)
- PointCNN++: Performant Convolution on Native Points (CVPR 2026 · similarity 0.89)
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.
