Research map / Point clouds / Segmentation and detection
Point-GNN: Graph Neural Network for 3D Object Detection in a Point Cloud
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 Point-GNN with point-cloud input | KITTI 3D object detection · KITTI test, Car Moderate | 3D AP: 77.63 -> 79.47 | Helped |
| Use Point-GNN with point-cloud input | KITTI 3D object detection · KITTI test, Cyclist Moderate | 3D AP: 62.53 -> 63.48 | Helped |
| Use Point-GNN with point-cloud input | KITTI 3D object detection · KITTI test, Pedestrian Moderate | 3D AP: 53.42 -> 43.77 | Hurt |
| Add auto-registration to standard NMS (ablation) | Point-GNN · KITTI val, Car Moderate | 3D AP: 76.80 -> 76.98 | Helped |
| Add box merging to auto-registration and standard NMS (ablation) | Point-GNN · KITTI val, Car Moderate | 3D AP: 76.98 -> 77.49 | Helped |
| Add box scoring to auto-registration and standard NMS (ablation) | Point-GNN · KITTI val, Car Moderate | 3D AP: 76.98 -> 77.90 | Helped |
| Combine box merging and scoring with auto-registration (ablation) | Point-GNN · KITTI val, Car Moderate | 3D AP: 78.40 -> 78.34 | Hurt |
| Increase graph message-passing iterations from zero to one (ablation) | Point-GNN · KITTI val, Car Moderate | 3D AP: 64.42 -> 77.89 | Helped |
| Increase graph message-passing iterations from two to three (ablation) | Point-GNN · KITTI val, Car Moderate | 3D AP: 78.51 -> 78.34 | Hurt |
| Reduce LiDAR scanning lines from 64 to 32 (ablation) | Point-GNN · KITTI val, Car Moderate | 3D AP: 78.34 -> 69.02 | Hurt |
| Reduce LiDAR scanning lines from 64 to 16 (ablation) | Point-GNN · KITTI val, Car Moderate | 3D AP: 78.34 -> 50.23 | Hurt |
| Reduce LiDAR scanning lines from 64 to 8 (ablation) | Point-GNN · KITTI val, Car Moderate | 3D AP: 78.34 -> 21.00 | Hurt |
Closest papers by meaning
- PC-RGNN: Point Cloud Completion and Graph Neural Network for 3D Object Detection (AAAI 2021 · similarity 0.92)
- VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection (CVPR 2018 · similarity 0.89)
- A Hierarchical Graph Network for 3D Object Detection on Point Clouds (CVPR 2020 · similarity 0.89)
- PointRCNN: 3D Object Proposal Generation and Detection From Point Cloud (CVPR 2019 · similarity 0.89)
- GBlobs: Explicit Local Structure via Gaussian Blobs for Improved Cross-Domain LiDAR-based 3D Object Detection (CVPR 2025 · similarity 0.86)
- Beyond Virtual Points: Depth-Enhanced LiDAR-only 3D Object Detection with Semi-Supervised Learning (Student Abstract) (AAAI 2025 · similarity 0.86)
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.