Research map / Label-efficient and robust learning / Domain adaptation and generalization
SiamDoGe: Domain Generalizable Semantic Segmentation Using Siamese Network
Zhenyao Wu, Xinyi Wu, Xiaoping Zhang, Lili Ju, Song Wang
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 second Siamese branch without consistency (ablation) | single-branch baseline · Cityscapes -> 5 targets (avg) | mIoU: 42.96 -> 43.30 | No clear effect |
| add second Siamese branch without consistency (ablation) | single-branch baseline · GTAV -> 5 targets (avg) | mIoU: 32.92 -> 32.83 | No clear effect |
| add sensitivity-guided consistency (ablation) | Siamese network · Cityscapes -> 5 targets (avg) | mIoU: 43.30 -> 45.42 | Improved |
| add sensitivity-guided consistency (ablation) | Siamese network · GTAV -> 5 targets (avg) | mIoU: 32.83 -> 34.00 | Improved |
| add self-guided randomization (ablation) | SiamDoGe with consistency · Cityscapes -> 5 targets (avg) | mIoU: 45.42 -> 46.92 | Improved |
| add self-guided randomization (ablation) | SiamDoGe with consistency · GTAV -> 5 targets (avg) | mIoU: 34.00 -> 35.75 | Improved |
| use MixStyle instead of self-guided randomization (ablation) | SiamDoGe · GTAV -> 5 targets (avg) | mIoU: 35.75 -> 34.15 | Hurt |
| randomize layer1 and layer2 instead of layer1 only (ablation) | self-guided randomization · GTAV -> 5 targets (avg) | mIoU: 35.23 -> 35.75 | Improved |
| randomize layer1 and layer2 instead of layer2 only (ablation) | self-guided randomization · Cityscapes -> 5 targets (avg) | mIoU: 45.30 -> 46.92 | Improved |
| add cropping operation C (ablation) | self-guided randomization · Cityscapes -> 5 targets (avg) | mIoU: 46.69 -> 46.92 | Improved |
| random lambda instead of fixed lambda = 0.5 (ablation) | self-guided randomization · GTAV -> 5 targets (avg) | mIoU: 34.60 -> 35.75 | Improved |
| set alpha to 10 instead of 1 (ablation) | consistency loss weight · GTAV -> 5 targets (avg) | mIoU: 33.67 -> 35.75 | Improved |
Closest papers by meaning
- A Simple Recipe for Language-guided Domain Generalized Segmentation (CVPR 2024 · similarity 0.89)
- Semantic-Aware Domain Generalized Segmentation (CVPR 2022 · similarity 0.89)
- Collaborating Foundation Models for Domain Generalized Semantic Segmentation (CVPR 2024 · similarity 0.87)
- Exploring Semantic Consistency and Style Diversity for Domain Generalized Semantic Segmentation (AAAI 2025 · similarity 0.87)
- FisherTune: Fisher-Guided Robust Tuning of Vision Foundation Models for Domain Generalized Segmentation (CVPR 2025 · similarity 0.87)
- Stronger Fewer & Superior: Harnessing Vision Foundation Models for Domain Generalized Semantic Segmentation (CVPR 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.
