Research map / Detection and segmentation / CNN features
An End-to-End Transformer Model for Crowd Localization
Dingkang Liang, Wei Xu, Xiang Bai
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 KMO (k-nearest-neighbour) term to Hungarian matching cost (ablation) | CLTR · UCF-QNRF | MAE: 91.3 -> 85.8 | Improved |
| add KMO term to Hungarian matching cost (ablation) | CLTR · UCF-QNRF | F1-measure: 80.02% -> 80.97% | Improved |
| use 500 queries instead of 300 (ablation) | CLTR · UCF-QNRF | MAE: 89.9 -> 85.8 | Improved |
| use 500 queries instead of 700 (ablation) | CLTR · UCF-QNRF | MAE: 86.8 -> 85.8 | Improved |
| use 6 layers instead of 3 (ablation) | CLTR · UCF-QNRF | MAE: 88.4 -> 85.8 | Improved |
| use 6 layers instead of 12 (ablation) | CLTR · UCF-QNRF | MAE: 87.7 -> 85.8 | Improved |
| use k=4 instead of k=3 (ablation) | CLTR · UCF-QNRF | MAE: 87.1 -> 85.8 | Improved |
| use k=4 instead of k=5 (ablation) | CLTR · UCF-QNRF | MAE: 86.9 -> 85.8 | Improved |
| use CLTR instead of GL on NWPU test counting | CLTR · NWPU-Crowd test | MAE: 79.3 -> 74.3 | Improved |
| use CLTR instead of AutoScale_loc on NWPU val localization | CLTR · NWPU-Crowd val | F1-measure: 66.8% -> 72.6% | Improved |
| use CLTR instead of TopoCount on UCF-QNRF counting | CLTR · UCF-QNRF | MAE: 89.0 -> 85.8 | No clear effect |
Closest papers by meaning
- Rethinking Counting and Localization in Crowds: A Purely Point-Based Framework (ICCV 2021 · similarity 0.87)
- NWPU-Crowd: A Large-Scale Benchmark for Crowd Counting and Localization (TPAMI 2020 · similarity 0.87)
- Local Spacing-Aware Hungarian Matching for Stable Point-Supervised Crowd Counting (ECCV 2026 · similarity 0.85)
- Locate, Size and Count: Accurately Resolving People in Dense Crowds via Detection (TPAMI 2020 · similarity 0.85)
- Point-to-Region Loss for Semi-Supervised Point-Based Crowd Counting (CVPR 2025 · similarity 0.82)
- Gramformer: Learning Crowd Counting via Graph-Modulated Transformer (AAAI 2024 · similarity 0.81)
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.
