Research map / Detection and segmentation / Weakly supervised detection and segmentation
H2FA R-CNN: Holistic and Hierarchical Feature Alignment for Cross-Domain Weakly Supervised Object Detection
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 DT+PL with H2FA R-CNN | Cross-domain weakly supervised object detection · Clipartall | mAP: 14.5% | Helped |
| Replace DT+PL with H2FA R-CNN | Cross-domain weakly supervised object detection · Cliparttest | mAP: 6.4% | Helped |
| Use H2FA R-CNN instead of the strongest compared method | Cross-domain weakly supervised object detection · Watercolor | mAP: ∼2% | Helped |
| Replace PLGE with H2FA R-CNN | Cross-domain weakly supervised object detection · Comic | mAP: 4.7% | Helped |
| Add image-level class-agnostic alignment (ablation) | H2FA R-CNN · Comic | mAP: +13.4% | Helped |
| Add image-level class-wise alignment (ablation) | H2FA R-CNN · Comic | mAP: +11.9% | Helped |
| Add instance-level foreground alignment (ablation) | H2FA R-CNN · Comic | mAP: +7.2% | Helped |
| Add instance-level class-wise alignment alone (ablation) | H2FA R-CNN · Cliparttest | mAP: -2.2% | Hurt |
| Add instance-level class-wise alignment after the other three alignments (ablation) | H2FA R-CNN · Clipartall | mAP: 63.1 (+33.9) -> 69.8 (+40.6) | Helped |
| Use all four hierarchical feature alignments instead of the source-only baseline (ablation) | H2FA R-CNN · Clipartall | mAP: +40.6 | Helped |
| Train with extra noisy target-domain samples | H2FA R-CNN · Watercolor | mAP: 2.7% | Helped |
| Apply H2FA R-CNN rather than the source-only baseline | Object detection adapted from CityScapes to Foggy Cityscapes · Foggy Cityscapes | mAP: +19.7% mAP on car | Helped |
Closest papers by meaning
- Cross-Domain Weakly-Supervised Object Detection Through Progressive Domain Adaptation (CVPR 2018 · similarity 0.89)
- Informative and Consistent Correspondence Mining for Cross-Domain Weakly Supervised Object Detection (CVPR 2021 · similarity 0.89)
- Cross-Domain Detection via Graph-Induced Prototype Alignment (CVPR 2020 · similarity 0.89)
- Universal Domain Adaptive Object Detection via Dual Probabilistic Alignment (AAAI 2025 · similarity 0.87)
- Mean Teacher DETR with Masked Feature Alignment: A Robust Domain Adaptive Detection Transformer Framework (AAAI 2024 · similarity 0.86)
- Improving Single Domain-Generalized Object Detection: A Focus on Diversification and Alignment (CVPR 2024 · similarity 0.86)
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.