Research map / 3D vision / Human pose estimation
Tracking by 3D Model Estimation of Unknown Objects in Videos
Denys Rozumnyi, Jiří Matas, Marc Pollefeys, Vittorio Ferrari, Martin R. Oswald
Read the paperCheck 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 |
|---|---|---|---|
| refine OSTrack segmentation with 3D model estimation | OSTrack · CDTB | IoU: 0.758 -> 0.788 | Improved |
| refine a weak older tracker (CSR-DCF) with the method | CSR-DCF · CDTB | IoU: 0.498 -> 0.512 | Improved |
| refine OSTrack on in-hand manipulation data | OSTrack · In-hand object manipulation | IoU (%): 93.26 -> 94.05 | Improved |
| use S2DNet features instead of RGB features (ablation) | Ours · Coin Tracking | IoU: 0.718 -> 0.740 | Improved |
| add flat-object prior (ablation) | Ours (S2DNet feat.) · Coin Tracking | IoU: 0.740 -> 0.764 | Improved |
| add feature rendering loss with S2DNet features (ablation) | Ours · CDTB (16 random sequences) | IoU: 0.727 -> 0.740 | Improved |
| add motion regularization L_M (ablation) | Ours · CDTB (16 random sequences) | IoU: 0.689 -> 0.740 | Improved |
| add Laplacian shape regularization L_L (ablation) | Ours · CDTB (16 random sequences) | IoU: 0.692 -> 0.740 | Improved |
| add texture total variation loss L_T (ablation) | Ours · CDTB (16 random sequences) | IoU: 0.734 -> 0.740 | Improved |
| use raw RGB features instead of dropping the feature loss (ablation) | Ours · CDTB (16 random sequences) | IoU: 0.727 -> 0.717 | Hurt |
| increase number of keyframes (ablation) | Ours · CDTB | average IoU | Improved |
Closest papers by meaning
- Seeing Behind Objects for 3D Multi-Object Tracking in RGB-D Sequences (CVPR 2021 · similarity 0.90)
- SpatialTracker: Tracking Any 2D Pixels in 3D Space (CVPR 2024 · similarity 0.89)
- A Discriminative Single-Shot Segmentation Network for Visual Object Tracking (TPAMI 2021 · similarity 0.89)
- Fast Spatial Tracking with Visual Geometry Transformer (CVPR 2026 · similarity 0.88)
- Prior-free 3D Object Tracking (CVPR 2025 · similarity 0.88)
- TrackCraft3R: Repurposing Video Diffusion Transformers for Dense 3D Tracking (NeurIPS 2026 · similarity 0.88)
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.
