Research map / Point clouds / Point cloud classification
Towards Efficient Graph Convolutional Networks for Point Cloud Handling
Yawei Li, He Chen, Zhaopeng Cui, Radu Timofte, Marc Pollefeys, Gregory S. Chirikjian, Luc Van Gool
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 |
|---|---|---|---|
| add KNN-search sharing with progressively enlarged neighbor pool (ablation) | DGCNN · ModelNet40 (2048 points, K=40) | inference time (ms): 385.8 -> 212.7 | Improved |
| add KNN-search sharing with progressively enlarged neighbor pool (ablation) | DGCNN · ModelNet40 (2048 points, K=40) | overall accuracy: 92.56 -> 92.58 | No clear effect |
| add shuffling of graph feature gathering and MLP (ablation) | DGCNN · ModelNet40 (2048 points, K=40) | inference time (ms): 385.8 -> 164.7 | Improved |
| add shuffling of graph feature gathering and MLP (ablation) | DGCNN · ModelNet40 (2048 points, K=40) | overall accuracy: 92.56 -> 92.63 | No clear effect |
| add both KNN sharing and operation shuffling | DGCNN · ModelNet40 (2048 points, K=40) | inference time (ms): 385.8 -> 132.0 | Improved |
| add both KNN sharing and operation shuffling | DGCNN · ModelNet40 (1024 points, K=20) | inference time (ms): 74.7 -> 38.1 | Improved |
| add both KNN sharing and operation shuffling | DGCNN · ModelNet40 (1024 points, K=20) | overall accuracy: 92.10 -> 92.56 | No clear effect |
| add both KNN sharing and operation shuffling | Third GCN · ModelNet40 (1024 points, K=20) | inference time (ms): 95.4 -> 24.1 | Improved |
| add both KNN sharing and operation shuffling | Third GCN · ModelNet40 (1024 points, K=20) | overall accuracy: 92.54 -> 92.50 | No clear effect |
| add both KNN sharing and operation shuffling | PointCNN · ModelNet40 (1024 points, K=20) | inference time (ms): 35.3 -> 29.1 | Improved |
| add both KNN sharing and operation shuffling | DGCNN · ShapeNetPart (2048 points) | mIoU: 84.95 -> 84.78 | No clear effect |
| add both KNN sharing and operation shuffling | DGCNN · S3DIS | mIoU: 57.5 -> 57.0 | Hurt |
Closest papers by meaning
- Convolution in the Cloud: Learning Deformable Kernels in 3D Graph Convolution Networks for Point Cloud Analysis (CVPR 2020 · similarity 0.92)
- Learning of 3D Graph Convolution Networks for Point Cloud Analysis (TPAMI 2021 · similarity 0.92)
- Long and Short-Range Dependency Graph Structure Learning Framework on Point Cloud (TPAMI 2023 · similarity 0.92)
- Grid-GCN for Fast and Scalable Point Cloud Learning (CVPR 2020 · similarity 0.91)
- Learning Adaptive Neighborhoods for Graph Neural Networks (ICCV 2023 · similarity 0.89)
- DHGCN: Dynamic Hop Graph Convolution Network for Self-Supervised Point Cloud Learning (AAAI 2024 · 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.
