Research map / 3D vision / Depth and stereo
View Birdification in the Crowd: Ground-Plane Localization from Perceived Movements
Mai Nishimura, Shohei Nobuhara, Ko Nishino
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 view birdification instead of plain Social Force extrapolation | Social Force · Hotel | absolute pedestrian error: 0.289 -> 0.052 | Improved |
| use view birdification instead of plain Social Force extrapolation | Social Force · ETH | absolute pedestrian error: 0.261 -> 0.079 | Improved |
| use view birdification instead of plain Social Force extrapolation | Social Force · Students | absolute pedestrian error: 0.222 -> 0.010 | Improved |
| use view birdification instead of ConstVel extrapolation | ConstVel · Shibuya01 | absolute pedestrian error: 0.221 -> 0.024 | Improved |
| use Social Force instead of constant velocity interaction model in VB | VB · ETH | absolute pedestrian error: 0.115 -> 0.079 | Improved |
| use Social Force instead of constant velocity interaction model in VB | VB · Students | absolute pedestrian error: 0.023 -> 0.010 | Improved |
| use Social Force instead of constant velocity interaction model in VB | VB · Hotel | absolute pedestrian error: 0.070 -> 0.052 | Improved |
| increase height variance sigma_h from 0.00 to 0.07 (ablation) | VB-SF · Hotel | absolute pedestrian error: 0.052 -> 0.052 | No clear effect |
| replace manual centerlines with MOT detections | VB · GTAV | absolute pedestrian error: 0.491 -> 0.530 | Hurt |
| replace manual centerlines with MOT detections | VB · GTAV | camera translation error: 0.097 -> 0.101 | Hurt |
| select only nearby pedestrians in super-dense crowd (ablation) | VB-SF · Shibuya01 | absolute pedestrian error: 0.023 | No clear effect |
| use nearby dynamic pedestrians instead of static keypoints beyond 6 m | VB vs five-point solver · GTAV simulation | camera translation error: 0.50 -> 0.097 | Improved |
Closest papers by meaning
- Delving Into the Devils of Bird’s-Eye-View Perception: A Review, Evaluation and Recipe (TPAMI 2023 · similarity 0.86)
- Crowd3D: Towards Hundreds of People Reconstruction From a Single Image (CVPR 2023 · similarity 0.85)
- CrowdDriven: A New Challenging Dataset for Outdoor Visual Localization (ICCV 2021 · similarity 0.85)
- DyCrowd: Towards Dynamic Crowd Reconstruction from a Large-scene Video (TPAMI 2025 · similarity 0.85)
- SeaBird: Segmentation in Bird's View with Dice Loss Improves Monocular 3D Detection of Large Objects (CVPR 2024 · similarity 0.84)
- GigaTraj: Predicting Long-term Trajectories of Hundreds of Pedestrians in Gigapixel Complex Scenes (CVPR 2024 · 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.
