Research map / Point clouds / Point cloud classification
CXTrack: Improving 3D Point Cloud Tracking With Contextual Information
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 |
|---|---|---|---|
| use CXTrack instead of STNet | 3D point-cloud tracking · KITTI Pedestrian | Success: 49.9 -> 67.0 | Helped |
| use CXTrack instead of STNet | 3D point-cloud tracking · KITTI Mean | Success: 61.3 -> 67.5 | Helped |
| use CXTrack instead of STNet | 3D point-cloud tracking · Waymo Open Dataset Pedestrian | Success: 25.5 -> 30.7 | Helped |
| use CXTrack instead of STNet | 3D point-cloud tracking · Waymo Open Dataset Vehicle | Success: 59.7 -> 57.1 | Hurt |
| use CXTrack instead of STNet | 3D point-cloud tracking · nuScenes Mean | Success: 26.9 -> 26.5 | Hurt |
| retain contextual information rather than cropping the previous-frame point cloud (ablation) | CXTrack · KITTI Mean | Success: 61.5 -> 67.5 | Helped |
| add cascaded targetness mask prediction (ablation) | CXTrack target-centric transformer · KITTI Mean | Success: 65.1 -> 67.5 | Helped |
| add auxiliary target center regression (ablation) | CXTrack target-centric transformer · KITTI Mean | Success: 56.6 -> 67.5 | Helped |
| use Semi-Dropout instead of Vanilla transformer layers (ablation) | CXTrack target-centric transformer · KITTI Mean | Success: 65.3 -> 67.5 | Helped |
| use Semi-Dropout instead of Gated transformer layers (ablation) | CXTrack target-centric transformer · KITTI Mean | Success: 64.1 -> 67.5 | Helped |
| add center embedding (ablation) | X-RPN · KITTI Mean | Success: 66.2 -> 67.5 | Helped |
| replace X-RPN with V2B localization head (ablation) | CXTrack · KITTI Pedestrian | Success: 60.1 -> 67.0 | Hurt |
Closest papers by meaning
- M3SOT: Multi-Frame, Multi-Field, Multi-Space 3D Single Object Tracking (AAAI 2024 · similarity 0.88)
- Towards Category Unification of 3D Single Object Tracking on Point Clouds (ICLR 2024 · similarity 0.87)
- Track-On: Transformer-based Online Point Tracking with Memory (ICLR 2025 · similarity 0.86)
- UAWTrack: Universal 3D Single Object Tracking in Adverse Weather (AAAI 2025 · similarity 0.86)
- Point Transformer V3: Simpler Faster Stronger (CVPR 2024 · similarity 0.85)
- TrackCraft3R: Repurposing Video Diffusion Transformers for Dense 3D Tracking (NeurIPS 2026 · similarity 0.85)
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.