Research map / Label-efficient and robust learning / Domain adaptation and generalization
Self-Ensembling Attention Networks: Addressing Domain Shift 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 self-ensembling attention domain adaptation vs no adaptation | DeepLab-v2 VGG-16 · SYNTHIA->Cityscapes | mIoU (%): 17.4 -> 37.5 | Helped |
| add self-ensembling attention domain adaptation vs no adaptation | DeepLab-v2 VGG-16 · GTA-5->Cityscapes | mIoU (%): 21.2 -> 35.7 | Helped |
| self-ensembling method vs CyCADA | domain adaptation · GTA-5->Cityscapes | mIoU (%): 34.8 -> 35.7 | Helped |
| increase consistency weight lambda_con from 0.3 to 10 (ablation) | self-ensembling attention network · SYNTHIA->Cityscapes | mIoU (%): 37.5 -> 21.4 | Hurt |
| increase consistency weight lambda_con from 0.3 to 3 (ablation) | self-ensembling attention network · SYNTHIA->Cityscapes | mIoU (%): 37.5 -> 24.2 | Hurt |
| lower lambda_con from 0.3 to 0.1 (ablation) | self-ensembling attention network · SYNTHIA->Cityscapes | mIoU (%): 37.5 -> 36.3 | No clear effect |
| raise attention threshold from 0.3 to 0.5 (ablation) | self-ensembling attention network · SYNTHIA->Cityscapes | mIoU (%): 37.5 -> 31.2 | Hurt |
| use attention threshold 0.3 vs no masking (0.0) (ablation) | self-ensembling attention network · SYNTHIA->Cityscapes | mIoU (%): 34.6 -> 37.5 | Helped |
Closest papers by meaning
- DAST: Unsupervised Domain Adaptation in Semantic Segmentation Based on Discriminator Attention and Self-Training (AAAI 2021 · similarity 0.91)
- Unsupervised Intra-Domain Adaptation for Semantic Segmentation Through Self-Supervision (CVPR 2020 · similarity 0.90)
- Self-Supervised Augmentation Consistency for Adapting Semantic Segmentation (CVPR 2021 · similarity 0.89)
- Bidirectional Learning for Domain Adaptation of Semantic Segmentation (CVPR 2019 · similarity 0.88)
- Balanced Learning for Domain Adaptive Semantic Segmentation (ICML 2025 · similarity 0.87)
- Pay Attention to Target: Relation-Aware Temporal Consistency for Domain Adaptive Video Semantic Segmentation (AAAI 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.