Research map / Label-efficient and robust learning / Recognition and long-tail
Residual Regression With Semantic Prior for 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 residual regression by comparing input images with labeled support images (ablation) | MCNN crowd counting · ShanghaiTech A | MAE: 86.97 -> 79.72 | Helped |
| Add residual regression by comparing input images with labeled support images (ablation) | CSRNet crowd counting · ShanghaiTech A | MAE: 68.2 -> 64.8 | Helped |
| Add semantic prior to appearance plus residual regression (ablation) | MCNN crowd counting · ShanghaiTech A | MAE: 79.72 -> 75.9 | Helped |
| Add semantic prior to appearance plus residual regression (ablation) | CSRNet crowd counting · ShanghaiTech A | MAE: 64.8 -> 64.2 | Helped |
| Add adversarial loss to the model with semantic prior (ablation) | MCNN crowd counting · ShanghaiTech A | MAE: 75.9 -> 72.6 | Helped |
| Add adversarial loss to the model with semantic prior (ablation) | CSRNet crowd counting · ShanghaiTech A | MAE: 64.2 -> 63.1 | Helped |
| Jointly optimize appearance and residual regression instead of using appearance alone (ablation) | MCNN trained on ShanghaiTech A · ShanghaiTech B (cross-dataset) | MAE: 44.7 -> 40.0 | Helped |
| Jointly optimize appearance and residual regression instead of using appearance alone (ablation) | MCNN trained on ShanghaiTech A · Expo (cross-dataset) | MAE: 62.2 -> 30.4 | Helped |
| Jointly optimize appearance and residual regression instead of using appearance alone (ablation) | MCNN trained on ShanghaiTech A · UCF 50 (cross-dataset) | MAE: 358.2 -> 355.0 | Helped |
| Use the proposed CSRNet framework rather than CSRNet as a published baseline | CSRNet · ShanghaiTech A | MAE: 68.2 -> 63.1 | Helped |
Closest papers by meaning
- Crowd Counting via Adversarial Cross-Scale Consistency Pursuit (CVPR 2018 · similarity 0.90)
- Context-Aware Crowd Counting (CVPR 2019 · similarity 0.90)
- Leveraging Heterogeneous Auxiliary Tasks to Assist Crowd Counting (CVPR 2019 · similarity 0.89)
- Switching Convolutional Neural Network for Crowd Counting (CVPR 2017 · similarity 0.89)
- Single Domain Generalization for Crowd Counting (CVPR 2024 · similarity 0.87)
- CrowdDiff: Multi-hypothesis Crowd Density Estimation using Diffusion Models (CVPR 2024 · 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.