Research map / Detection and segmentation / Open-set and few-shot detection
Adaptive Object Detection with Dual Multi-Label Prediction
Zhen Zhao, Yuhong Guo, Haifeng Shen, Jieping Ye
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 multi-label conditional adversary (MC) instead of no alignment (ablation) | Faster R-CNN · Cityscapes to Foggy Cityscapes | mAP (%): 25.1 -> 36.6 | Improved |
| condition the adversary on multi-label predictions instead of unconditional adversary (ablation) | MCAR · Cityscapes to Foggy Cityscapes | mAP (%): 33.7 -> 36.6 | Improved |
| add prediction consistency regularization (PR) (ablation) | MCAR · Cityscapes to Foggy Cityscapes | mAP (%): 36.6 -> 38.8 | Improved |
| add prediction consistency regularization (PR) (ablation) | MCAR with unconditional adversary · Cityscapes to Foggy Cityscapes | mAP (%): 31.0 -> 33.7 | Improved |
| add multi-label prediction loss (ablation) | MCAR with unconditional adversary without PR · Cityscapes to Foggy Cityscapes | mAP (%): 30.5 -> 31.0 | Improved |
| add prediction regularization to MC | MCAR (MC only) · PASCAL VOC to Watercolor | mAP (%): 54.4 -> 56.0 | Improved |
| add prediction regularization to MC | MCAR (MC only) · PASCAL VOC to Comic | mAP (%): 32.6 -> 33.5 | Improved |
| apply MCAR domain adaptation instead of source-only | Source-only Faster R-CNN · PASCAL VOC to Watercolor | mAP (%): 44.6 -> 56.0 | Improved |
| apply MCAR instead of source-only | Source-only Faster R-CNN · PASCAL VOC to Comic | mAP (%): 19.7 -> 33.5 | Improved |
| use MCAR instead of SCL | SCL · Cityscapes to Foggy Cityscapes | mAP (%): 37.9 -> 38.8 | Improved |
| reduce focal modulation gamma from 5 to 1 (ablation) | MCAR adversary · PASCAL VOC to Watercolor (hyperparameter sensitivity) | mAP (%): 54.4 -> 44.0 | Hurt |
| increase focal modulation gamma from 5 to 9 (ablation) | MCAR adversary · PASCAL VOC to Watercolor (hyperparameter sensitivity) | mAP (%): 54.4 -> 44.8 | Hurt |
Closest papers by meaning
- Domain Adaptive Object Detection via Balancing Between Self-Training and Adversarial Learning (TPAMI 2023 · similarity 0.89)
- Adapting Object Detectors via Selective Cross-Domain Alignment (CVPR 2019 · similarity 0.89)
- Domain Adaptive Object Detection via Asymmetric Tri-way Faster-RCNN (ECCV 2020 · similarity 0.89)
- Robust Domain Adaptive Object Detection With Unified Multi-Granularity Alignment (TPAMI 2024 · similarity 0.89)
- Exploring Categorical Regularization for Domain Adaptive Object Detection (CVPR 2020 · similarity 0.89)
- Unified Domain Generalization and Adaptation for Multi-View 3D Object Detection (NeurIPS 2024 · similarity 0.87)
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.
