Research map / Detection and segmentation / CNN features
Pedestrian Detection with Spatially Pooled Features and Structured Ensemble Learning
Sakrapee Paisitkriangkrai, Chunhua Shen, Anton van den Hengel
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 spatial pooling to covariance features (ablation) | sp-Cov · INRIA | log-average miss rate (%): 14.2% -> 12.8% | Improved |
| add spatial pooling to covariance features (ablation) | sp-Cov · ETH | log-average miss rate (%): 42.7% -> 42.0% | Improved |
| add spatial pooling to covariance features (ablation) | sp-Cov · TUD-Brussels | log-average miss rate (%): 48.6% -> 47.8% | Improved |
| add spatial pooling to LBP features (ablation) | sp-LBP · INRIA | log-average miss rate (%): 23.2% -> 21.8% | Improved |
| add spatial pooling to LBP features (ablation) | sp-LBP · ETH | log-average miss rate (%): 47.2% -> 47.1% | No clear effect |
| add spatial pooling to LBP features (ablation) | sp-LBP · TUD-Brussels | log-average miss rate (%): 54.9% -> 54.4% | Improved |
| add M+O+LUV channels and sp-LBP to sp-Cov (ablation) | sp-Cov · TUD-Brussels | log-average miss rate (%): 47.8% -> 42.5% | Improved |
| add M+O+LUV channels to sp-Cov+LUV (ablation) | sp-Cov · INRIA | log-average miss rate (%): 12.8% -> 11.2% | Improved |
| use sp-Cov+LUV instead of M+O+LUV+LBP | sp-Cov · ETH | log-average miss rate (%): 39.9% -> 42.0% | Hurt |
| use the structured ensemble learner instead of AdaBoost | pAUC ensemble · FACE (100 weak classifiers) | average pAUC [0,0.1]: 0.90 -> 0.91 | Improved |
| use the structured ensemble learner instead of AdaBoost | pAUC ensemble · SCENE (100 weak classifiers) | average pAUC [0,0.1]: 0.85 -> 0.87 | Improved |
| use the structured ensemble learner instead of AsymBoost | pAUC ensemble · USPS (100 weak classifiers) | average pAUC [0,0.1]: 0.96 -> 0.97 | Improved |
Closest papers by meaning
- How Far Are We From Solving Pedestrian Detection? (CVPR 2016 · similarity 0.86)
- Towards Reaching Human Performance in Pedestrian Detection (TPAMI 2017 · similarity 0.86)
- Pedestrian Detection With Autoregressive Network Phases (CVPR 2019 · similarity 0.85)
- Generalizable Pedestrian Detection: The Elephant in the Room (CVPR 2021 · similarity 0.85)
- Multi-Label Learning of Part Detectors for Heavily Occluded Pedestrian Detection (ICCV 2017 · similarity 0.85)
- Learning Efficient Single-stage Pedestrian Detectors by Asymptotic Localization Fitting (ECCV 2018 · 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.
