Research map / 3D vision / Depth and stereo
CSRNet: Dilated Convolutional Neural Networks for Understanding the Highly Congested Scenes
Read the paperCheck 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 |
|---|---|---|---|
| Replace CP-CNN with CSRNet | crowd counting · ShanghaiTech Part_A | MAE: 73.6 -> 68.2 | Helped |
| Replace MCNN with CSRNet | crowd counting · ShanghaiTech Part_B | MAE: 26.4 -> 10.6 | Helped |
| Replace CP-CNN with CSRNet | crowd counting · UCF_CC_50 | MAE: 295.8 -> 266.1 | Helped |
| Replace CP-CNN with CSRNet | crowd counting · WorldExpo'10 | Avg. MAE: 8.86 -> 8.6 | Helped |
| Replace MCNN with CSRNet | crowd counting · UCSD | MAE: 1.07 -> 1.16 | Hurt |
| Replace FCN-HA with CSRNet | vehicle counting · TRANCOS | GAME(0): 4.21 -> 3.56 | Helped |
| Use dilation rate 2 rather than rate 1 throughout the back-end (ablation) | CSRNet · ShanghaiTech Part_A | MAE: 69.7 -> 68.2 | Helped |
| Use dilation rates 2 then 4 rather than rate 2 throughout the back-end (ablation) | CSRNet · ShanghaiTech Part_A | MAE: 68.2 -> 71.91 | Hurt |
| Use dilation rate 4 rather than rate 2 throughout the back-end (ablation) | CSRNet · ShanghaiTech Part_A | MAE: 68.2 -> 75.81 | Hurt |
| Use dropout (ablation) | CSRNet · ShanghaiTech Part_A | MAE | No clear effect |
| Use a deeper single-column CNN instead of MCNN | crowd counting · ShanghaiTech Part_A | MAE: 110.2 -> 93.0 | Helped |
Closest papers by meaning
- Understanding Traffic Density From Large-Scale Web Camera Data (CVPR 2017 · similarity 0.84)
- Densely Connected Convolutional Networks (CVPR 2017 · similarity 0.84)
- Spatial as Deep: Spatial CNN for Traffic Scene Understanding (AAAI 2018 · similarity 0.83)
- DenseASPP for Semantic Segmentation in Street Scenes (CVPR 2018 · similarity 0.83)
- Dilated Residual Networks (CVPR 2017 · similarity 0.83)
- Recurrent Partial Kernel Network for Efficient Optical Flow Estimation (AAAI 2024 · similarity 0.83)
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.