Research map / Video / Video segmentation and tracking
ORBIT: Benchmarking SfM in the Wild with 360deg Video
Sara Sabour, Richard Tucker, Marcus Brubaker, Saurabh Saxena, Junhwa Hur, Andrea Tagliasacchi, Deqing Sun, David J. Fleet, Richard Szeliski, Noah Snavely
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 |
|---|---|---|---|
| use MegaSaM instead of COLMAP | SfM on ORBIT · ORBIT | failure rate (%): 56.49 -> 39.09 | Improved |
| add RoMo motion masks to MegaSaM (ablation) | MegaSaM · ORBIT | mean ATE: 0.67 -> 0.59 | Improved |
| add RoMo motion masks to MegaSaM (ablation) | MegaSaM · ORBIT low-texture clips | median ATE: 0.58 -> 0.78 | Hurt |
| use feed-forward VGGT-Long or MonST3R instead of bundle-adjustment methods | SfM on ORBIT · ORBIT | failure rate (%): 94.72 and 98.44 vs 39.09 (MegaSaM) | Hurt |
| use VGGT-Long instead of COLMAP | SfM on ORBIT · ORBIT | mean ATE: mean ATE 1.34 -> 1.30; failure rate 56.49 -> 94.72 | Mixed |
| lack of texture in static regions | COLMAP and MegaSaM · ORBIT low-texture clips | median ATE: COLMAP 1.59; MegaSaM 0.58 | Hurt |
| high camera speed | VGGT-Long · ORBIT high-speed clips | median ATE: 1.61 | Hurt |
| use MegaSaM instead of COLMAP on crowded scenes | SfM on ORBIT · ORBIT crowded clips | median ATE: 0.43 -> 0.08 | Improved |
| use 360-degree rig-based SfM to create ground truth | ORBIT pipeline · 360Loc | ATE (m): 0.07 +- 0.04 m, orders of magnitude below baseline failures | Improved |
Closest papers by meaning
- OmniLife360: A Benchmark for 3D Reconstruction from In-the-Wild 360° Captures (ECCV 2026 · similarity 0.89)
- VGGSfM: Visual Geometry Grounded Deep Structure From Motion (CVPR 2024 · similarity 0.88)
- VideoSfM: Exploiting Temporal Structure for Video-Based Structure-from-Motion (ECCV 2026 · similarity 0.88)
- MP-SfM: Monocular Surface Priors for Robust Structure-from-Motion (CVPR 2025 · similarity 0.87)
- ProxyPose: 6-DoF Pose Tracking via Video-to-Video Translation (NeurIPS 2026 · similarity 0.87)
- TAPVid-360: Tracking Any Point in 360 from Narrow Field of View Video (NeurIPS 2025 · similarity 0.86)
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.
