Research map / Detection and segmentation / Open-set and few-shot detection
Open Compound Domain Adaptation with Object Style Compensation for Semantic Segmentation
Tingliang Feng, Hao Shi, Xueyang Liu, Wei Feng, Liang Wan, Yanlin Zhou, Di Lin
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 OLDM discrepancy memorization and compensation to segmentation backbone (ablation) | OLDM object style compensation · GTA5->C-Driving | mIoU(T): 36.6 -> 44.1 | Improved |
| add OLDM (open domain) (ablation) | OLDM object style compensation · GTA5->C-Driving open | mIoU(O): 39.7 -> 46.9 | Improved |
| use global rather than local category-key features (ablation) | OLDM object style compensation · GTA5->C-Driving | mIoU(T): 41.7 -> 44.1 | Improved |
| use category-key instead of mean of instance-key features (ablation) | OLDM object style compensation · GTA5->C-Driving | mIoU(T): 39.2 -> 44.1 | Improved |
| update 100% of instance-key features instead of top-1 (ablation) | OLDM object style compensation · GTA5->C-Driving | mIoU(T): 40.7 -> 44.1 | Improved |
| update 100% rather than top-50% instance-key features (ablation) | OLDM object style compensation · GTA5->C-Driving | mIoU(T): 42.2 -> 44.1 | Improved |
| use instance-key similarity weighting instead of mean discrepancy (ablation) | OLDM object style compensation · GTA5->C-Driving | mIoU(T): 37.8 -> 44.1 | Improved |
| use category-separated multi-sets instead of merged set (ablation) | OLDM object style compensation · GTA5->C-Driving | mIoU(T): 39.6 -> 44.1 | Improved |
| use category-based multi-sets instead of k-means (ablation) | OLDM object style compensation · GTA5->C-Driving | mIoU(T): 41.7 -> 44.1 | Improved |
| use discrepancy features instead of key discrepancy (ablation) | OLDM object style compensation · GTA5->C-Driving | mIoU(T): 42.7 -> 44.1 | Improved |
| weight discrepancy by instance similarity at test time instead of averaging (ablation) | OLDM object style compensation · GTA5->C-Driving | mIoU(T): 40.7 -> 44.1 | Improved |
| add pseudo annotations from final compensated features (ablation) | OLDM object style compensation · GTA5->C-Driving | mIoU(T): 39.7 -> 44.1 | Improved |
Closest papers by meaning
- Style Mixing and Patchwise Prototypical Matching for One-Shot Unsupervised Domain Adaptive Semantic Segmentation (AAAI 2022 · similarity 0.90)
- Discover, Hallucinate, and Adapt: Open Compound Domain Adaptation for Semantic Segmentation (NeurIPS 2020 · similarity 0.89)
- Pulling Target to Source: A New Perspective on Domain Adaptive Semantic Segmentation (IJCV 2024 · similarity 0.89)
- Open-Set Domain Adaptation for Semantic Segmentation (CVPR 2024 · similarity 0.89)
- On the Road to Online Adaptation for Semantic Image Segmentation (CVPR 2022 · similarity 0.89)
- Universal Domain Adaptation for Semantic Segmentation (CVPR 2025 · 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.
