Research map / 3D vision / Depth and stereo
Cross-View Cross-Scene Multi-View Crowd Counting
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 camera selection and noise-view regularization (ablation) | Backbone crowd counter · Synthetic CVCS test | MAE: 14.13 -> 7.22 | Helped |
| Add camera selection (ablation) | Backbone crowd counter · Synthetic CVCS test | MAE: 14.13 -> 8.63 | Helped |
| Add noise-view regularization (ablation) | Backbone crowd counter · Synthetic CVCS test | MAE: 14.13 -> 7.94 | Helped |
| Replace the backbone fusion with MVMS (ablation) | Backbone crowd counter · Synthetic CVCS test | MAE: 14.13 -> 9.30 | Helped |
| Use a 3-layer CNN rather than no convolution for camera selection (ablation) | Backbone+CamSel · Synthetic CVCS test | MAE: 10.77 -> 8.15 | Helped |
| Inject noise directly into the input image (ablation) | Backbone crowd counter · Synthetic CVCS test | MAE: 14.13 -> 19.91 | Hurt |
| Add dropout after feature extractor layers 2 and 4 (ablation) | Backbone crowd counter · Synthetic CVCS test | MAE: 14.13 -> 13.16 | Helped |
| Pass the noise view through the shared feature extractor and projection before max fusion (ablation) | Backbone crowd counter · Synthetic CVCS test | MAE: 14.13 -> 7.94 | Helped |
| Use sum rather than max to fuse projected noise features (ablation) | Backbone+NoiseV · Synthetic CVCS test | MAE: 7.94 -> 8.57 | Hurt |
| Use a separate noise feature extractor rather than the shared extractor (ablation) | Backbone+NoiseV · Synthetic CVCS test | MAE: 7.94 -> 8.48 | Hurt |
| Increase input cameras from 5 to 11 (ablation) | CVCS · Synthetic CVCS test | MAE: 7.22 -> 7.00 | Helped |
| Apply unsupervised domain adaptation using unlabeled target-scene images (ablation) | CVCS trained on synthetic data · PETS2009 | MAE: 5.33 -> 5.17 | Helped |
Closest papers by meaning
- Wide-Area Crowd Counting via Ground-Plane Density Maps and Multi-View Fusion CNNs (CVPR 2019 · similarity 0.90)
- 3D Crowd Counting via Multi-View Fusion with 3D Gaussian Kernels (AAAI 2020 · similarity 0.90)
- Switching Convolutional Neural Network for Crowd Counting (CVPR 2017 · similarity 0.88)
- Leveraging Self-Supervision for Cross-Domain Crowd Counting (CVPR 2022 · similarity 0.88)
- Multi-view Crowd Tracking Transformer with View-Ground Interactions Under Large Real-World Scenes (CVPR 2026 · similarity 0.87)
- Embodied Crowd Counting (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.