Research map / Label-efficient and robust learning / Domain adaptation and generalization
WildNet: Learning Domain Generalized Semantic Segmentation From the Wild
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Extend source content and style using WildNet | ResNet-50 segmentation model trained on GTAV · Average over C, B, M, S, G | mIoU: 39.06 -> 46.33 | Helped |
| Add CEL (ablation) | GTAV-trained ResNet-50 baseline · Average over C, B, M, S, G | mIoU: 39.06 -> 43.46 | Helped |
| Add SEL to CEL (ablation) | GTAV-trained ResNet-50 with CEL · Average over C, B, M, S, G | mIoU: 43.46 -> 45.48 | Helped |
| Add SCR to CEL and SEL (ablation) | GTAV-trained ResNet-50 with CEL and SEL · Average over C, B, M, S, G | mIoU: 45.48 -> 46.33 | Helped |
| Use more wild images for training (ablation) | WildNet · Average over C, B, M, S, G | mIoU: 44.60 -> 45.31 -> 45.45 -> 45.59 -> 46.33 | Helped |
| Apply FS to residual groups 1–3 rather than extending into groups 4–5 (ablation) | WildNet · Average over C, B, M, S, G | mIoU: 46.33 -> 45.19 -> 44.34 | Hurt |
| Increase wild-content dictionary size (ablation) | WildNet · Average over C, B, M, S, G | mIoU: 44.98 -> 45.27 -> 45.73 -> 46.33 | Helped |
| Use wild feature statistics rather than random source-statistic scaling for FS (ablation) | WildNet · Average over C, B, M, S, G | mIoU: 43.54 -> 46.33 | Helped |
| Use uniform rather than random CEL feature-map sampling (ablation) | WildNet · Average over C, B, M, S, G | mIoU: 45.89 -> 46.33 | Helped |
| Extend source contents and styles to wild examples instead of relying on style whitening | Domain-generalized semantic segmentation · Cityscapes, BDD100K, Mapillary, SYNTHIA | mIoU: 38.78 -> 44.62; 35.64 -> 38.42; 40.38 -> 46.09; 28.97 -> 31.34 | Helped |
Closest papers by meaning
- Exploring Semantic Consistency and Style Diversity for Domain Generalized Semantic Segmentation (AAAI 2025 · similarity 0.90)
- A Simple Recipe for Language-guided Domain Generalized Segmentation (CVPR 2024 · similarity 0.89)
- Collaborating Foundation Models for Domain Generalized Semantic Segmentation (CVPR 2024 · similarity 0.89)
- SemiDAViL: Semi-supervised Domain Adaptation with Vision-Language Guidance for Semantic Segmentation (CVPR 2025 · similarity 0.88)
- Balanced Learning for Domain Adaptive Semantic Segmentation (ICML 2025 · similarity 0.88)
- CLIPCEIL: Domain Generalization through CLIP via Channel rEfinement and Image-text aLignment (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.