Research map / Point clouds / Registration and matching
Improving Point-based Crowd Counting and Localization Based on Auxiliary Point Guidance
I-HSIANG CHEN, Wei-Ting Chen, Yu-Wei Liu, Ming-Hsuan Yang, Sy-Yen Kuo
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 |
|---|---|---|---|
| replace nearest-neighbor feature lookup with full IFI (ablation) | APGCC · ShanghaiTech A | MAE: 53.16 -> 48.84 | Improved |
| replace bilinear interpolation with IFI (ablation) | APGCC · ShanghaiTech A | MAE: 51.25 -> 48.84 | Improved |
| add positional encoding to IFI distances (ablation) | IFI · ShanghaiTech A | MAE: 49.24 -> 48.84 | Improved |
| use four nearest features instead of a single reference point in IFI (ablation) | IFI · ShanghaiTech A | MAE: 49.73 -> 48.84 | Improved |
| use nearest-point proposal selection instead of Matcher (ablation) | point-based counting · ShanghaiTech A | count error | Hurt |
| combine Matcher with APG instead of APG alone (ablation) | APGCC · ShanghaiTech A | MAE | Improved |
| increase auxiliary point pairs from (2,2) to (5,5) (ablation) | APG · ShanghaiTech A | MAE | No clear effect |
| add auxiliary negative points to positive ones (ablation) | APG · ShanghaiTech A | training stability | Improved |
| apply APGCC versus P2PNet for counting | APGCC · ShanghaiTech A | MAE: 52.7 -> 48.8 | Improved |
| apply APGCC for localization on ShanghaiTech A | APGCC · ShanghaiTech A | F1-measure: +5.5% at sigma=4; +3.8% at sigma=8 | Improved |
Closest papers by meaning
- Rethinking Counting and Localization in Crowds: A Purely Point-Based Framework (ICCV 2021 · similarity 0.89)
- Point-to-Region Loss for Semi-Supervised Point-Based Crowd Counting (CVPR 2025 · similarity 0.88)
- Recurrent Attentive Zooming for Joint Crowd Counting and Precise Localization (CVPR 2019 · similarity 0.87)
- Robust Zero-Shot Crowd Counting and Localization with Adaptive Resolution SAM (ECCV 2024 · similarity 0.87)
- Point-Query Quadtree for Crowd Counting, Localization, and More (ICCV 2023 · similarity 0.87)
- Embodied Crowd Counting (NeurIPS 2025 · 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.
