Research map / Architectures and efficiency / Attention and transformers
Boosting Crowd Counting via Multifaceted Attention
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 |
|---|---|---|---|
| Use MAN instead of the strongest reported baseline, P2PNet | Crowd counting models · ShanghaiTech A | MAE: 52.7 -> 56.8 | Hurt |
| Use MAN instead of the strongest reported baseline, S3 | Crowd counting models · UCF-QNRF | MAE: 80.6 -> 77.3 | Helped |
| Use MAN instead of S3 | Crowd counting models · JHU++ | MAE: 59.4 -> 53.4 | Helped |
| Use MAN instead of the strongest reported baseline, P2PNet | Crowd counting models · NWPU | MAE: 77.4 -> 76.5 | Helped |
| Add a vanilla Transformer encoder (ablation) | BL baseline · UCF-QNRF | MAE: 88.7 -> 85.2 | Helped |
| Add Instance Attention Loss (ablation) | BL baseline · UCF-QNRF | MAE: 88.7 -> 84.7 | Helped |
| Add Learnable Region Attention (ablation) | BL baseline · UCF-QNRF | MAE: 88.7 -> 84.2 | Helped |
| Add IAL to the Transformer (ablation) | Transformer crowd counter · UCF-QNRF | MAE: 85.2 -> 83.0 | Helped |
| Add LRA to the Transformer (ablation) | Transformer crowd counter · UCF-QNRF | MAE: 85.2 -> 82.9 | Helped |
| Add LRA to the Transformer (ablation) | Transformer crowd counter · UCF-QNRF | MSE: 149.5 -> 144.2 | Helped |
| Add LAR to Transformer + LRA + IAL (ablation) | Transformer + LRA + IAL · UCF-QNRF | MAE: 81.5 -> 77.3 | Helped |
| Add IAL to Transformer + LRA + LAR (ablation) | Transformer + LRA + LAR · UCF-QNRF | MAE: 80.5 -> 77.3 | Helped |
Closest papers by meaning
- Attention Scaling for Crowd Counting (CVPR 2020 · similarity 0.91)
- Gramformer: Learning Crowd Counting via Graph-Modulated Transformer (AAAI 2024 · similarity 0.90)
- Shallow Feature Based Dense Attention Network for Crowd Counting (AAAI 2020 · similarity 0.89)
- Context-Aware Crowd Counting (CVPR 2019 · similarity 0.89)
- Switching Convolutional Neural Network for Crowd Counting (CVPR 2017 · similarity 0.88)
- Free Lunch Enhancements for Multi-modal Crowd Counting (CVPR 2025 · similarity 0.85)
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.