Research map / 3D vision / Depth and stereo
Real-Time 3D Model Tracking in Color and Depth on a Single CPU Core
Wadim Kehl, Federico Tombari, Slobodan Ilic, Nassir Navab
Check your idea against 137,217 papers
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 |
|---|---|---|---|
| Add cloud-based weighting to the color posterior (ablation) | Ours (A), joint contour and cloud tracking · Choi dataset | translation error (mm): 0.58 -> 0.51 | Improved |
| Add cloud-based weighting to the color posterior (ablation) | Ours (A), joint contour and cloud tracking · Choi dataset | rotation error (degrees): 0.28 -> 0.26 | Improved |
| Add cloud-based weighting to the color posterior (ablation) | Ours (A), joint contour and cloud tracking · Choi dataset | runtime (ms): 2.73 -> 6.92 | Hurt |
| Use joint contour and cloud tracking with cloud weighting instead of Tan et al. | Choi dataset tracking · Choi dataset | translation error (mm): 0.51 vs 0.51 | Improved |
| Use joint contour and cloud tracking with cloud weighting instead of Tan et al. | Choi dataset tracking · Choi dataset | rotation error (degrees): 0.26 vs 0.26 | Improved |
| Use joint contour and cloud tracking with cloud weighting instead of Tan et al. | Choi dataset tracking · Choi dataset | runtime (ms): 1.5 vs 6.92 | Hurt |
| Rely on contour energy alone rather than combining contour and ICP (ablation) | Joint contour and cloud tracking · Kinect Box and Tide sequences | translational error: λ = 10 | Hurt |
| Rely on plane-to-point ICP energy alone rather than combining contour and ICP (ablation) | Joint contour and cloud tracking · Kinect Box sequence | translational error: λ = 109 | Hurt |
| Balance contour and ICP terms with λ = 105 (ablation) | Joint contour and cloud tracking · Kinect Box sequence | translational error: λ = 105 | Improved |
| Vary the number of contour and interior sample points (ablation) | Joint contour and cloud tracking · Tide sequence | translation and rotation error: 10; 30; 80-90 | Mixed |
Closest papers by meaning
- Real-Time Tracking of Single and Multiple Objects from Depth-Colour Imagery Using 3D Signed Distance Functions (IJCV 2017 · similarity 0.88)
- TrackingWorld: World-centric Monocular 3D Tracking of Almost All Pixels (NeurIPS 2025 · similarity 0.85)
- DELTA: DENSE EFFICIENT LONG-RANGE 3D TRACKING FOR ANY VIDEO (ICLR 2025 · similarity 0.85)
- SpatialTracker: Tracking Any 2D Pixels in 3D Space (CVPR 2024 · similarity 0.85)
- TAPIP3D: Tracking Any Point in Persistent 3D Geometry (NeurIPS 2025 · similarity 0.85)
- TrackCraft3R: Repurposing Video Diffusion Transformers for Dense 3D Tracking (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.
