Research map / Learning theory and algorithms / Prediction and decision losses
Discrete-Constrained Regression for Local Counting Models
Haipeng Xiong, Angela Yao
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 standard regression with discrete-constrained regression | regression · ShanghaiTech A | MAE: 65.4 -> 61.6 | Improved |
| replace standard regression with discrete-constrained regression | regression · UCF-QNRF | MAE: 98.6 -> 91.4 | Improved |
| add global count loss L_bias^lambda (ablation) | DC-regression · ShanghaiTech A | MAE: 61.6 -> 60.3 | Improved |
| add global count loss L_BL (ablation) | DC-regression · ShanghaiTech A | MAE: 61.6 -> 60.7 | Improved |
| add naive global count loss L_c (ablation) | DC-regression · ShanghaiTech A | MAE: 61.6 -> 63.5 | Hurt |
| add L_bias^0 (no threshold lambda) (ablation) | DC-regression · ShanghaiTech A | MAE: 61.6 -> 61.6 | No clear effect |
| add global count loss L_BL (ablation) | regression · ShanghaiTech A | MAE: 65.4 -> 62.1 | Improved |
| use log instead of linear intervals (ablation) | DC-regression · ShanghaiTech A | MAE: 62.5 -> 61.6 | Improved |
| use log instead of linear intervals (ablation) | classification · ShanghaiTech A/B | MAE: SHA 65.6 -> 64.3, SHB 8.6 -> 8.7 | Mixed |
| replace classification with DC-regression in S-DCNet | S-DCNet · JHU-Crowd / UCF-QNRF / ShanghaiTech A | MAE: SHA 58.3 -> 59.8, JHU 65.2 -> 62.1, QNRF 104.4 -> 84.8 | Mixed |
| use DC-regression instead of regression | regression · MegaAge | MAE: 5.26 -> 5.15 | Improved |
| use DC-regression instead of regression | regression · MegaAsian | MAE: 2.87 -> 2.80 | Improved |
Closest papers by meaning
- Adaptive Mixture Regression Network with Local Counting Map for Crowd Counting (ECCV 2020 · similarity 0.88)
- Counting without Scale: Scale-Consistent Error Correction for Crowd Counting (NeurIPS 2026 · similarity 0.88)
- Distribution Matching for Crowd Counting (NeurIPS 2020 · similarity 0.86)
- Single Domain Generalization for Crowd Counting (CVPR 2024 · similarity 0.86)
- Point in, Box Out: Beyond Counting Persons in Crowds (CVPR 2019 · similarity 0.86)
- CrowdDiff: Multi-hypothesis Crowd Density Estimation using Diffusion Models (CVPR 2024 · 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.
