Research map / 3D vision / Depth and stereo
CountFormer: Multi-View Crowd Counting Transformer
Hong Mo, Xiong Zhang, Jianchao Tan, Cheng Yang, Qiong Gu, Bo Hang, Wenqi Ren
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 |
|---|---|---|---|
| replace IPM with deformable-attention feature lifting module (ablation) | CountFormer · PETS2009 | MAE: 2.76 -> 1.31 | Improved |
| replace IPM with feature lifting module (ablation) | CountFormer · CVCS | MAE: 11.3 -> 9.51 | Improved |
| add volume-level camera embedding (ablation) | CountFormer · CVCS | MAE: 9.51 -> 6.22 | Improved |
| add multi-view volume aggregation module instead of averaging (ablation) | CountFormer · CVCS | MAE: 6.22 -> 4.96 | Improved |
| add multi-view volume aggregation module (ablation) | CountFormer · PETS2009 | MAE: 1.17 -> 0.80 | Improved |
| add camera encoding to image-level features (ablation) | CountFormer · CVCS | MAE: 4.96 -> 4.79 | Improved |
| add camera encoding to image-level features (ablation) | CountFormer · CityStreet | NAE: 0.057 -> 0.058 | No clear effect |
| add feature lifting module to CVCS method (ablation) | CVCS baseline · CVCS | MAE: 7.27 -> 7.41 | No clear effect |
| add L, V, A modules to existing MVC methods (ablation) | 3D Counting · PETS2009 | MAE: 3.25 -> 1.29 | Improved |
| add volume camera embedding and aggregation to BEVFormer (ablation) | BEVFormer · CVCS | MAE: 9.56 -> 6.33 | Improved |
| train with dynamic number of views U(2,11) instead of fixed (ablation) | CountFormer · CVCS, 3 test views | MAE: 26.8 -> 24.0 | Improved |
| train with U(1,3) views instead of 3 views (ablation) | CountFormer · PETS2009, 1 test view | MAE: 7.05 -> 6.01 | Improved |
Closest papers by meaning
- 3D Crowd Counting via Geometric Attention-Guided Multi-view Fusion (IJCV 2022 · similarity 0.89)
- 3D Crowd Counting via Multi-View Fusion with 3D Gaussian Kernels (AAAI 2020 · similarity 0.88)
- Calibration-Free Multi-View Crowd Counting (ECCV 2022 · similarity 0.88)
- Wide-Area Crowd Counting: Multi-view Fusion Networks for Counting in Large Scenes (IJCV 2022 · similarity 0.88)
- Cross-View Cross-Scene Multi-View Crowd Counting (CVPR 2021 · similarity 0.87)
- Wide-Area Crowd Counting via Ground-Plane Density Maps and Multi-View Fusion CNNs (CVPR 2019 · similarity 0.87)
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.
