Research map / Label-efficient and robust learning / Domain adaptation and generalization
Addressing Domain Gap via Content Invariant Representation for Semantic Segmentation
Read the paperCheck 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 |
|---|---|---|---|
| Add a content-invariant intermediate domain (CIR) to source-only training (ablation) | CIRN · GTAV → Cityscapes | mIoU: 36.6 -> 42.7 | Helped |
| Add adaptive semantic segmentation to source-only training (ablation) | CIRN · GTAV → Cityscapes | mIoU: 36.6 -> 43.6 | Helped |
| Add both CIR and adaptive semantic segmentation (ablation) | CIRN · GTAV → Cityscapes | mIoU: 43.6 -> 46.1 | Helped |
| Add self-training with pseudo labels to CIR and adaptive semantic segmentation (ablation) | CIRN · GTAV → Cityscapes | mIoU: 46.1 -> 49.1 | Helped |
| Add a content-invariant intermediate domain (CIR) to source-only training (ablation) | CIRN · SYNTHIA → Cityscapes | mIoU: 33.5 -> 39.7 | Helped |
| Add adaptive semantic segmentation to source-only training (ablation) | CIRN · SYNTHIA → Cityscapes | mIoU: 33.5 -> 39.5 | Helped |
| Add both CIR and adaptive semantic segmentation (ablation) | CIRN · SYNTHIA → Cityscapes | mIoU: 39.5 -> 41.5 | Helped |
| Add self-training with pseudo labels to CIR and adaptive semantic segmentation (ablation) | CIRN · SYNTHIA → Cityscapes | mIoU: 41.5 -> 43.9 | Helped |
| Use ACM with two attention modules instead of a normal classifier (K=0) (ablation) | CIRN stage 2 · GTAV → Cityscapes | mIoU: 44.5 -> 46.1 | Helped |
| Use ACM with two attention modules instead of a normal classifier (K=0) (ablation) | CIRN stage 2 · SYNTHIA → Cityscapes | mIoU: 39.9 -> 41.5 | Helped |
| Increase ACM attention modules from K=2 to K=3 (ablation) | CIRN stage 2 · GTAV → Cityscapes | mIoU: 46.1 -> 45.1 | Hurt |
| Increase ACM attention modules from K=2 to K=3 (ablation) | CIRN stage 2 · SYNTHIA → Cityscapes | mIoU: 41.5 -> 41.7 | Helped |
Closest papers by meaning
- All About Structure: Adapting Structural Information Across Domains for Boosting Semantic Segmentation (CVPR 2019 · similarity 0.92)
- Unsupervised Intra-Domain Adaptation for Semantic Segmentation Through Self-Supervision (CVPR 2020 · similarity 0.91)
- Semantic-Aware Domain Generalized Segmentation (CVPR 2022 · similarity 0.90)
- Universal Domain Adaptation for Semantic Segmentation (CVPR 2025 · similarity 0.90)
- Balanced Learning for Domain Adaptive Semantic Segmentation (ICML 2025 · similarity 0.89)
- Open-Set Domain Adaptation for Semantic Segmentation (CVPR 2024 · similarity 0.89)
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.