Research map / 3D vision / Depth and stereo
Human Detection and Segmentation via Multi-View Consensus
Isinsu Katircioglu, Helge Rhodin, Jörg Spörri, Mathieu Salzmann, Pascal Fua
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 center and height consistency (3D projection of boxes) beyond consistent sampling (ablation) | Multi-view consensus detector · Ski-PTZ (4 cameras) | F score: 0.75 -> 0.83 | Improved |
| add center and height consistency beyond consistent sampling (ablation) | Multi-view consensus detector · H36M (4 cameras) | mAP@0.5: 0.79 -> 0.85 | Improved |
| add bounding box height consistency (ablation) | Multi-view consensus detector · Ski-PTZ (4 cameras) | F score: 0.79 -> 0.83 | Improved |
| add bounding box height consistency (ablation) | Multi-view consensus detector · Ski-PTZ (4 cameras) | J score: 0.70 -> 0.71 | Improved |
| add bounding box height consistency (ablation) | Multi-view consensus detector · H36M (4 cameras) | mAP@0.5: 0.83 -> 0.85 | Improved |
| use projection-based view consistency instead of triangulation loss (ablation) | Multi-view consensus detector · Ski-PTZ (4 cameras) | F score: 0.75 -> 0.83 | Improved |
| use projection-based view consistency instead of triangulation loss (ablation) | Multi-view consensus detector · H36M (4 cameras) | mAP@0.5: 0.82 -> 0.85 | Improved |
| add width consistency to full model (ablation) | Multi-view consensus detector · Ski-PTZ (4 cameras) | F score: 0.83 -> 0.77 | Hurt |
| add width consistency to full model (ablation) | Multi-view consensus detector · H36M (4 cameras) | mAP@0.5: 0.85 -> 0.84 | Hurt |
| increase training cameras from 4 to 6 (ablation) | Multi-view consensus detector · Ski-PTZ | F score: 0.83 -> 0.76 | Hurt |
| increase 3D grid from 10^3 to 16^3 voxels (ablation) | Multi-view consensus detector · Ski-PTZ | J score: 0.64 -> 0.68 | Improved |
| increase 3D grid from 10^3 to 16^3 voxels (ablation) | Multi-view consensus detector · H36M | mAP@0.5: 0.79 -> 0.76 | Hurt |
Closest papers by meaning
- Self-Supervised Human Detection and Segmentation via Background Inpainting (TPAMI 2021 · similarity 0.88)
- Towards Generalizable Multi-Camera 3D Object Detection via Perspective Rendering (AAAI 2025 · similarity 0.87)
- Self-Supervised Object Detection from Egocentric Videos (ICCV 2023 · similarity 0.87)
- Learning Accurate Segmentation Purely from Self-Supervision (ECCV 2026 · similarity 0.86)
- Tracking-by-Detection of 3D Human Shapes: From Surfaces to Volumes (TPAMI 2017 · similarity 0.86)
- Unified Domain Generalization and Adaptation for Multi-View 3D Object Detection (NeurIPS 2024 · similarity 0.85)
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.
