Research map / Point clouds / Point cloud classification
Graph-Based Point Tracker for 3D Object Tracking in Point Clouds
Read the paperCheck 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 |
|---|---|---|---|
| replace P2B-style feature fusion with bipartite graph feature augmentation and ITC | GPT · KITTI tracking, Car | Success: 56.2 -> 59.1 | Helped |
| replace P2B-style feature fusion with bipartite graph feature augmentation and ITC | GPT · KITTI tracking, Pedestrian | Precision: 49.6 -> 63.6 | Helped |
| add ITC using template-center voting features (ablation) | GPT · KITTI Car | Success: 58.9 -> 59.1 | Helped |
| add ITC using template-center voting features (ablation) | GPT · KITTI Car | Precision: 75.3 -> 75.6 | Helped |
| use separate feature transforms for template and search area instead of no transform (ablation) | GPT GFA · KITTI Car | Success: 56.9 -> 58.9 | Helped |
| share feature-transform weights between template and search area (ablation) | GPT GFA · KITTI Car | Success: 56.9 -> 57.0 | No clear effect |
| set graph neighbor count to k=16 rather than k=1 (ablation) | GPT GFA · KITTI Car | Success: 55.6 -> 59.1 | Helped |
| use first-frame GT plus previous tracking result to build the template rather than first-frame GT alone (ablation) | GPT · KITTI Car | Success: 52.6 -> 59.1 | Helped |
Closest papers by meaning
- P2B: Point-to-Box Network for 3D Object Tracking in Point Clouds (CVPR 2020 · similarity 0.87)
- Point-GNN: Graph Neural Network for 3D Object Detection in a Point Cloud (CVPR 2020 · similarity 0.87)
- GNN3DMOT: Graph Neural Network for 3D Multi-Object Tracking With 2D-3D Multi-Feature Learning (CVPR 2020 · similarity 0.86)
- Learnable Graph Matching: Incorporating Graph Partitioning With Deep Feature Learning for Multiple Object Tracking (CVPR 2021 · similarity 0.86)
- Towards Category Unification of 3D Single Object Tracking on Point Clouds (ICLR 2024 · similarity 0.83)
- M3SOT: Multi-Frame, Multi-Field, Multi-Space 3D Single Object Tracking (AAAI 2024 · similarity 0.83)
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.