Research map / 3D vision / Depth and stereo
CarFusion: Combining Point Tracking and Part Detection for Dynamic 3D Reconstruction of Vehicles
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 |
|---|---|---|---|
| Fuse structured part detections with unstructured feature tracks (ablation) | CarFusion · Straight | reprojection RMSE: 7.1 -> 2.5 | Helped |
| Fuse structured part detections with unstructured feature tracks (ablation) | CarFusion · Turning | reprojection RMSE: 5.83 -> 3.1 | Helped |
| Fuse structured part detections with unstructured feature tracks (ablation) | CarFusion · Multi | reprojection RMSE: 4.47 -> 2.2 | Helped |
| Fine-tune the detector using reprojected 3D part locations (ablation) | CarFusion · Intersection Dataset | PCK (α = 0.1): 87.1 -> 91.4 | Helped |
| Fine-tune the detector using reprojected 3D part locations (ablation) | CarFusion · Intersection Dataset | PCK (α = 0.2): 91.8 -> 94.5 | Helped |
| Use temporal integration of structured detections without feature-track fusion (ablation) | T-cRANSAC · Straight | reprojection RMSE: 12.24 -> 17.8; 8.52 -> 7.1 | Mixed |
| Use temporal integration of structured detections without feature-track fusion (ablation) | T-cRANSAC · Turning | reprojection RMSE: 8.94 -> 12.5; 6.95 -> 5.83 | Mixed |
| Use temporal integration of structured detections without feature-track fusion (ablation) | T-cRANSAC · Multi | reprojection RMSE: 7.45 -> 14.3; 5.3 -> 4.47 | Mixed |
| Apply multiview bootstrapping to the detector (ablation) | cRANSAC · Straight | reprojection RMSE: 12.24 -> 8.52 | Helped |
| Apply multiview bootstrapping to the detector (ablation) | T-cRANSAC · Straight | reprojection RMSE: 17.8 -> 7.1 | Helped |
| Apply multiview bootstrapping to the detector (ablation) | CarFusion · Straight | reprojection RMSE: 16.8 -> 2.5 | Helped |
Closest papers by meaning
- M3SOT: Multi-Frame, Multi-Field, Multi-Space 3D Single Object Tracking (AAAI 2024 · similarity 0.86)
- WALT3D: Generating Realistic Training Data from Time-Lapse Imagery for Reconstructing Dynamic Objects Under Occlusion (CVPR 2024 · similarity 0.86)
- FRUC: Feedforward Dynamic Scene Reconstruction from Uncalibrated Collaborative Driving Views (NeurIPS 2026 · similarity 0.85)
- 3D StreetUnveiler with Semantic-aware 2DGS - a simple baseline (ICLR 2025 · similarity 0.85)
- CorrBEV: Multi-View 3D Object Detection by Correlation Learning with Multi-modal Prototypes (CVPR 2025 · similarity 0.85)
- Effective Multi-sensor Conditioning for Street-view Novel-view Synthesis (NeurIPS 2026 · similarity 0.84)
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.