Research map / 3D vision / Depth and stereo
3D Random Occlusion and Multi-layer Projection for Deep Multi-Camera Pedestrian Localization
Rui Qiu, Ming Xu, Yuyao Yan, Jeremy S. Smith, Xi Yang
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 |
|---|---|---|---|
| add multi-layer projection (ablation) | MVDet · MultiviewX | MODA: 0.839 -> 0.900 | Improved |
| add multi-layer projection (ablation) | MVDet · WILDTRACK | MODA: 0.882 -> 0.912 | Improved |
| add multi-layer projection (ablation) | MVDet · Terrace | MODA: 0.872 -> 0.894 | Improved |
| add 3D Random Occlusion (ablation) | MVDet · MultiviewX | MODA: 0.839 -> 0.898 | Improved |
| add 3D Random Occlusion (ablation) | MVDet · WILDTRACK | MODA: 0.882 -> 0.923 | Improved |
| add 3D Random Occlusion (ablation) | MVDet · Terrace | MODA: 0.872 -> 0.915 | Improved |
| add 3D Random Occlusion on top of multi-layer projection (ablation) | 3DROM · MultiviewX | MODA: 0.900 -> 0.950 | Improved |
| add multi-layer projection on top of 3D Random Occlusion (ablation) | 3DROM · MultiviewX | MODA: 0.898 -> 0.950 | Improved |
| replace per-view Random Erasing with 3D Random Occlusion (ablation) | 3DROM · MultiviewX | MODA: 0.927 -> 0.950 | Improved |
| replace per-view Random Erasing with 3D Random Occlusion (ablation) | 3DROM · Terrace | MODA: 0.923 -> 0.948 | Improved |
| raise occlusion probability p from 0% to 100% (ablation) | 3DROM · WILDTRACK | MODA: 0.882 -> 0.935 | Improved |
| use extra plane at 180 cm (0,180 vs single 0 cm plane) (ablation) | 3DROM · Terrace | MODA: 0.915 -> 0.867 | Hurt |
Closest papers by meaning
- Stacked Homography Transformations for Multi-View Pedestrian Detection (ICCV 2021 · similarity 0.85)
- Where, What, Whether: Multi-Modal Learning Meets Pedestrian Detection (CVPR 2020 · similarity 0.84)
- DCHM: Depth-Consistent Human Modeling for Multiview Detection (ICCV 2025 · similarity 0.84)
- Bi-box Regression for Pedestrian Detection and Occlusion Estimation (ECCV 2018 · similarity 0.83)
- Occlusion-aware R-CNN: Detecting Pedestrians in a Crowd (ECCV 2018 · similarity 0.83)
- Multi-View Pedestrian Occupancy Prediction with a Novel Synthetic Dataset (AAAI 2025 · similarity 0.82)
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.
