Research map / Label-efficient and robust learning / Recognition and long-tail
Crowd Counting in the Frequency Domain
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 |
|---|---|---|---|
| replace spatial density-map loss with characteristic-function loss | VGG19 crowd-counting model · NWPU | MAE: 79.3 -> 76.8 | Helped |
| replace spatial density-map loss with characteristic-function loss | VGG19 crowd-counting model · JHU++ | MAE: 59.9 -> 57.0 | Helped |
| replace spatial density-map loss with characteristic-function loss | VGG19 crowd-counting model · UCF-QNRF | MAE: 84.3 -> 80.3 | Helped |
| replace spatial density-map loss with characteristic-function loss | VGG19 crowd-counting model · ShanghaiTech A | MAE: 61.3 -> 57.5 | Helped |
| replace spatial density-map loss with characteristic-function loss | VGG19 crowd-counting model · ShanghaiTech B | MAE: 7.3 -> 6.9 | Helped |
| use ChfL instead of the strongest reported single-model baseline | VGG19 crowd-counting model · ShanghaiTech A and ShanghaiTech B | MAE: 52.7 -> 57.5; 6.2 -> 6.9 | Hurt |
| set integral range to [-0.1, 0.1]2 rather than the chosen [-0.3, 0.3]2 (ablation) | ChfL · ShanghaiTech A | MAE: 57.5 -> 58.9 | Hurt |
| set integral range to [-0.2, 0.2]2 rather than the chosen [-0.3, 0.3]2 (ablation) | ChfL · ShanghaiTech A | MAE: 57.5 -> 59.0 | No clear effect |
| expand integral range beyond [-0.3, 0.3]2 (ablation) | ChfL · ShanghaiTech A | MAE: 57.5 -> 60.6; 57.5 -> 61.7 | Hurt |
| use coarse Riemann grid granularity 0.1 (ablation) | ChfL · ShanghaiTech A | MAE: 59.0 -> 60.3 | Hurt |
| vary grid granularity below 0.04 (ablation) | ChfL · ShanghaiTech A | MAE: 59.0; 59.0; 58.6; 59.6 | No clear effect |
Closest papers by meaning
- Modeling Noisy Annotations for Crowd Counting (NeurIPS 2020 · similarity 0.88)
- Domain-General Crowd Counting in Unseen Scenarios (AAAI 2023 · similarity 0.88)
- 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)
- Free Lunch Enhancements for Multi-modal Crowd Counting (CVPR 2025 · similarity 0.86)
- Gaussian Density Splatting Network (NeurIPS 2026 · similarity 0.84)
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.