Research map / Label-efficient and robust learning / Recognition and long-tail
Crowd Counting With Deep Structured Scale Integration Network
Lingbo Liu, Zhilin Qiu, Guanbin Li, Shufan Liu, Wanli Ouyang, Liang Lin
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 CRF-based structured feature enhancement (2 iterations) instead of no feature refinement (ablation) | DSSINet · ShanghaiTech Part A | MAE: 68.85 -> 60.63 | Improved |
| use CRF refinement instead of concatenation + 1x1 conv (ablation) | DSSINet · ShanghaiTech Part A | MAE: 67.11 -> 60.63 | Improved |
| increase CRF iterations from 2 to 3 (ablation) | DSSINet · ShanghaiTech Part A | MAE: 60.63 -> 63.80 | Hurt |
| increase CRF iterations from 1 to 2 (ablation) | DSSINet · ShanghaiTech Part A | MAE: 64.37 -> 60.63 | Improved |
| use image pyramid with 3 scales instead of single scale (ablation) | DSSINet · ShanghaiTech Part A | MAE: 70.94 -> 60.63 | Improved |
| add a fourth pyramid scale (0.25) (ablation) | DSSINet · ShanghaiTech Part A | MAE: 60.63 -> 61.13 | No clear effect |
| use DMS-SSIM-5 loss instead of Euclidean loss (ablation) | DSSINet · ShanghaiTech Part A | MAE: 67.68 -> 60.63 | Improved |
| use DMS-SSIM-5 loss instead of plain SSIM loss (ablation) | DSSINet · ShanghaiTech Part A | MAE: 74.60 -> 60.63 | Improved |
| add dilation to multiscale SSIM loss (DMS-SSIM-5 vs MS-SSIM-5) (ablation) | DSSINet · ShanghaiTech Part A | MAE: 63.51 -> 60.63 | Improved |
| add a sixth dilated layer (rate 9) to DMS-SSIM loss (ablation) | DSSINet · ShanghaiTech Part A | MAE: 60.63 -> 62.60 | Hurt |
| increase DMS-SSIM scales from 2 to 5 (ablation) | DSSINet · ShanghaiTech Part A | MAE: 73.33 -> 60.63 | Improved |
Closest papers by meaning
- Multi-Level Bottom-Top and Top-Bottom Feature Fusion for Crowd Counting (ICCV 2019 · similarity 0.91)
- Scale Aggregation Network for Accurate and Efficient Crowd Counting (ECCV 2018 · similarity 0.91)
- To Choose or to Fuse? Scale Selection for Crowd Counting (AAAI 2021 · similarity 0.90)
- Shallow Feature Based Dense Attention Network for Crowd Counting (AAAI 2020 · similarity 0.90)
- Counting without Scale: Scale-Consistent Error Correction for Crowd Counting (NeurIPS 2026 · similarity 0.90)
- Context-Aware Crowd Counting (CVPR 2019 · similarity 0.90)
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.
