Research map / Detection and segmentation / Open-set and few-shot detection
Bidirectional Alignment for Domain Adaptive Detection with Transformers
Liqiang He, Wei Wang, Albert Chen, Min Sun, Cheng-Hao Kuo, Sinisa Todorovic
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 BiADT bidirectional domain-invariant/specific alignment over AQT | DAB-Deformable-DETR · Cityscapes->FoggyCityscapes | mAP50: 47.4 -> 49.4 | Improved |
| add object-level alignments to backbone alignment (ablation) | BiADT · Cityscapes->FoggyCityscapes | mAP50: 38.3 -> 42.9 | Improved |
| add image-level alignments to backbone alignment (ablation) | BiADT · Cityscapes->FoggyCityscapes | mAP50: 38.3 -> 44.4 | Improved |
| combine object and image alignments (ablation) | BiADT · Cityscapes->FoggyCityscapes | mAP50: 44.4 -> 48.2 | Improved |
| add DyHinge loss (ablation) | BiADT · Cityscapes->FoggyCityscapes | mAP50: 48.2 -> 48.7 | Improved |
| add mutual information loss (ablation) | BiADT · Cityscapes->FoggyCityscapes | mAP50: 48.7 -> 49.4 | Improved |
| add AQT alignments to BiADT | BiADT · Cityscapes->FoggyCityscapes | mAP50: 49.4 -> 49.8 | Improved |
| integrate BiADT into teacher-student framework with pseudo-labels | BiADT · Cityscapes->FoggyCityscapes | mAP50: 49.4 -> 50.8 | Improved |
| use Conditional-DETR instead of Dab-Deformable-DETR detector (ablation) | BiADT · Cityscapes->FoggyCityscapes | mAP50: 49.4 -> 27.8 | Hurt |
| use Dab-Deformable-DETR instead of Deformable-DETR (ablation) | BiADT · Cityscapes->FoggyCityscapes | mAP50: 48.9 -> 49.4 | Improved |
| compare BiADT with pseudo-label methods | BiADT · Cityscapes->FoggyCityscapes | mAP50: 49.4 vs AT 50.9 | Mixed |
Closest papers by meaning
- Adapting Object Detectors via Selective Cross-Domain Alignment (CVPR 2019 · similarity 0.90)
- Domain Adaptive Object Detection via Asymmetric Tri-way Faster-RCNN (ECCV 2020 · similarity 0.89)
- iFAN: Image-Instance Full Alignment Networks for Adaptive Object Detection (AAAI 2020 · similarity 0.89)
- Multi-Adversarial Faster-RCNN for Unrestricted Object Detection (ICCV 2019 · similarity 0.89)
- Universal Domain Adaptive Object Detection via Dual Probabilistic Alignment (AAAI 2025 · similarity 0.88)
- Mean Teacher DETR with Masked Feature Alignment: A Robust Domain Adaptive Detection Transformer Framework (AAAI 2024 · similarity 0.88)
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.
