Research map / Detection and segmentation / Open-set and few-shot detection
Boosting Domain Generalized and Adaptive Detection with Diffusion Models: Fitness, Generalization, and Transferability
Boyong He, Yuxiang Ji, Zhuoyue Tan, Liaoni Wu
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 consistency loss between auxiliary and ordinary branch (ablation) | Diff. detector · FoggyCityscapes | mAP: 45.8 -> 50.7 | Improved |
| add consistency loss between auxiliary and ordinary branch (ablation) | Diff. detector · Cityscapes in-domain | mAP: 62.4 -> 61.1 | Hurt |
| add noise adding in single-step diffusion (ablation) | Diff. detector · Foggy. cross-domain | mAP: 30.8 -> 34.9 | Improved |
| add feature collection and fusion (ablation) | Diff. detector · Cityscapes in-domain | mAP: 38.5 -> 52.4 | Improved |
| add object-centered auxiliary branch (ablation) | Diff. detector · Clipart | mAP: 43.6 -> 53.3 | Improved |
| use single-step diffusion instead of five steps | GDD · Cityscapes | mAP: 59.8 -> 36.8 | Hurt |
| vary consistency weight gamma from 0.0 to 1.0 (ablation) | Diff. detector · FoggyCityscapes | mAP: 47.6 -> 50.7 | Improved |
| increase gamma from 1.0 to 2.0 (ablation) | Diff. detector · FoggyCityscapes | mAP: 50.7 -> 49.1 | Hurt |
| add feature alignment on top of domain augmentation (ablation) | Diff. guided detector DG · BDD100K | mAP: 36.2 -> 39.5 | Improved |
| add object alignment (pseudo-labels) for DG guided detector (ablation) | Diff. guided detector DG · BDD100K | mAP: 39.5 -> 46.7 | Improved |
| add object alignment for DA guided detector (ablation) | Diff. guided detector DA · BDD100K | mAP: 40.2 -> 51.3 | Improved |
| add domain augmentation (ablation) | Diff. guided detector DG · BDD100K | mAP: 25.4 -> 36.2 | Improved |
Closest papers by meaning
- Generalized Diffusion Detector: Mining Robust Features from Diffusion Models for Domain-Generalized Detection (CVPR 2025 · similarity 0.96)
- Towards Single-Source Domain Generalized Object Detection via Causal Visual Prompts (NeurIPS 2025 · similarity 0.89)
- SEEN-DA: SEmantic ENtropy guided Domain-aware Attention for Domain Adaptive Object Detection (CVPR 2025 · similarity 0.88)
- Towards Unsupervised Model Selection for Domain Adaptive Object Detection (NeurIPS 2024 · similarity 0.88)
- Unbiased Faster R-CNN for Single-source Domain Generalized Object Detection (CVPR 2024 · similarity 0.88)
- Object-Aware Domain Generalization for Object Detection (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.
