Research map / Vision-language and multimodal / CLIP and zero-shot
UPRE: Zero-Shot Domain Adaptation for Object Detection via Unified Prompt and Representation Enhancement
Xiao Zhang, Fei Wei, Yong Wang, Wenda Zhao, Feiyi Li, Xiangxiang Chu
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 learnable prompt to complete prompt (ablation) | UPRE · Diverse Weather | mAP: 37.2 -> 40.0 (Daytime Foggy) | Improved |
| use shared learnable context instead of separate (ablation) | UPRE · Night Rainy | mAP: 19.8 -> 17.4 | Hurt |
| add L_s to L_a in RDD (ablation) | UPRE · Diverse Weather | mAP: 29.7 -> 31.3 (Dusk Rainy), MAD 0.8 -> 1.7 | Mixed |
| add L_r to RDD (ablation) | UPRE · Diverse Weather | mAP: 31.3 -> 32.5 (Dusk Rainy) | Improved |
| add domain adaptation prompt (ablation) | UPRE · Daytime Foggy | mAP: 37.8 -> 40.0 | Improved |
| add unified representation enhancement (ablation) | UPRE · Daytime Foggy | mAP: 38.5 -> 40.0 | Improved |
| add relative domain distance (Img) (ablation) | UPRE · Daytime Foggy | mAP: 39.2 -> 40.0 | Improved |
| add positive-negative separation (Ins) (ablation) | UPRE · Daytime Foggy | mAP: 38.9 -> 40.0 | Improved |
| add enhancement and Img/Ins to prompt-only training (ablation) | UPRE · Daytime Foggy | mAP: 32.1 -> 40.0 | Improved |
| region-level (MxN) enhancement instead of image-level 1x1 (ablation) | UPRE · Diverse Weather | mAP: 36.7 -> 40.0 (Daytime Foggy) | Improved |
| add mean enhancement E_mu on top of E_sigma (ablation) | UPRE · Diverse Weather | mAP: 38.0 -> 40.0 (Daytime Foggy, MxN) | Improved |
| use unbiased prompt prior instead of Gaussian noise (ablation) | UPRE · Daytime Foggy | mAP: 34.5 -> 40.0 | Improved |
Closest papers by meaning
- Towards Single-Source Domain Generalized Object Detection via Causal Visual Prompts (NeurIPS 2025 · similarity 0.88)
- DA-Ada: Learning Domain-Aware Adapter for Domain Adaptive Object Detection (NeurIPS 2024 · similarity 0.88)
- Zero-shot Generalizable Incremental Learning for Vision-Language Object Detection (NeurIPS 2024 · similarity 0.87)
- Domain-Agnostic Mutual Prompting for Unsupervised Domain Adaptation (CVPR 2024 · similarity 0.87)
- Unified Language-driven Zero-shot Domain Adaptation (CVPR 2024 · similarity 0.87)
- Style Evolving along Chain-of-Thought for Unknown-Domain Object Detection (CVPR 2025 · 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.
