Research map / Point clouds / Point cloud classification
Synchronize Feature Extracting and Matching: A Single Branch Framework for 3D Object Tracking
Teli Ma, Mengmeng Wang, Jimin Xiao, Huifeng Wu, Yong Liu
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 |
|---|---|---|---|
| replace Siamese backbone plus matcher (STNet matcher) with single-branch synchronized extraction and matching (ablation) | SyncTrack · KITTI Car | Success: 71.8 -> 73.3 | Improved |
| replace Siamese backbone plus V2B matcher with single-branch structure (ablation) | SyncTrack · nuScenes Car | Success: 35.3 -> 36.7 | Improved |
| add attentive sampling on search region tokens (vs random sampling) (ablation) | APST · KITTI Car | Success: 70.9 -> 73.3 | Improved |
| add attentive sampling on search region tokens (vs FPS) (ablation) | APST · KITTI Car | Success: 71.1 -> 73.2 | Improved |
| apply attentive sampling to template tokens only (ablation) | APST · KITTI Car | Success: 70.9 -> 67.6 | Hurt |
| apply attentive sampling to both template and search tokens (ablation) | APST · KITTI Car | Success: 73.3 -> 69.6 | Hurt |
| increase backbone depth and width (Small to Large) (ablation) | SyncTrack · nuScenes Bicycle | Success: 23.8 -> 25.6 | Improved |
| increase backbone depth and width (Small to Large) (ablation) | SyncTrack · nuScenes Truck | Success: 39.4 -> 40.5 | Improved |
| use single-branch SyncTrack instead of M2Track | 3D SOT · KITTI Pedestrian | Success: 61.5 -> 54.7 | Hurt |
| use SyncTrack instead of STNet on mean results | 3D SOT · KITTI (Mean) | Success: 61.3 -> 64.1 | Improved |
Closest papers by meaning
- Beyond 3D Siamese Tracking: A Motion-Centric Paradigm for 3D Single Object Tracking in Point Clouds (CVPR 2022 · similarity 0.90)
- An Effective Motion-Centric Paradigm for 3D Single Object Tracking in Point Clouds (TPAMI 2023 · similarity 0.89)
- 3D Siamese Transformer Network for Single Object Tracking on Point Clouds (ECCV 2022 · similarity 0.89)
- TrackAny3D: Transferring Pretrained 3D Models for Category-unified 3D Point Cloud Tracking (ICCV 2025 · similarity 0.89)
- 3D Siamese Voxel-to-BEV Tracker for Sparse Point Clouds (NeurIPS 2021 · similarity 0.88)
- Modeling Continuous Motion for 3D Point Cloud Object Tracking (AAAI 2024 · similarity 0.87)
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.
