Research map / Point clouds / Shape and surface completion
Leveraging Shape Completion for 3D Siamese Tracking
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 |
|---|---|---|---|
| Add shape-completion loss to tracking-loss training (ablation) | 3D Siamese tracker · KITTI Tracking, OPE | Success: 73.96 -> 76.94 | Helped |
| Add shape-completion loss to tracking-loss training (ablation) | 3D Siamese tracker · KITTI Tracking, OPE | Precision: 78.68 -> 81.38 | Helped |
| Train with completion loss only instead of tracking loss only (ablation) | 3D Siamese tracker · KITTI Tracking, OPE | Success: 73.96 -> 65.36 | Hurt |
| Train with completion loss only instead of tracking loss only (ablation) | 3D Siamese tracker · KITTI Tracking, OPE | Precision: 78.68 -> 70.62 | Hurt |
| Pre-train the network on ShapeNet (ablation) | 3D Siamese tracker · KITTI Tracking, OPE | Success: 39.06 -> 44.54 | Helped |
| Use first and previous point clouds rather than all previous point clouds (ablation) | 3D Siamese tracker with early fusion · KITTI Tracking, OPE | Precision: 81.4 -> 82.7 | Helped |
| Use max pooling rather than median pooling for latent fusion (ablation) | 3D Siamese tracker with latent fusion · KITTI Tracking, OPE | Success: 59.7 -> 71.5 | Helped |
| Use early point-cloud fusion rather than late latent fusion (ablation) | 3D Siamese tracker · KITTI Tracking, OPE | Success: 76.9 / 81.4 vs 63.9 / 73.2 | Mixed |
| Use a Kalman Filter to propose candidates instead of exhaustive search | 3D Siamese tracker · KITTI Tracking test, OPE3D | Success: 76.94 -> 40.09 | Hurt |
| Use 3D Siamese tracking with exhaustive search rather than AVOD tracking-by-detection | KITTI Tracking · KITTI Tracking test, OPE3D | Success: 63.16 -> 76.94 | Helped |
| Add shape-completion loss to tracking-loss training (ablation) | 3D Siamese tracker · KITTI Tracking, fully visible and occluded samples | Success: 76.9/80.1 -> 79.9/83.2; 72.8/77.4 -> 74.6/80.5 | Helped |
| Use symmetry as a prior for shape completion (ablation) | 3D Siamese tracker · KITTI Tracking | tracking performance | No clear effect |
Closest papers by meaning
- Robust 3D Tracking with Quality-Aware Shape Completion (AAAI 2024 · similarity 0.89)
- Towards Category Unification of 3D Single Object Tracking on Point Clouds (ICLR 2024 · similarity 0.88)
- Modeling Continuous Motion for 3D Point Cloud Object Tracking (AAAI 2024 · similarity 0.86)
- Scaling Diffusion Models to Real-World 3D LiDAR Scene Completion (CVPR 2024 · similarity 0.86)
- Beyond Virtual Points: Depth-Enhanced LiDAR-only 3D Object Detection with Semi-Supervised Learning (Student Abstract) (AAAI 2025 · similarity 0.86)
- Towards Learning to Complete Anything in Lidar (ICML 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.