Research map / Label-efficient and robust learning / Domain adaptation and generalization
Significance-Aware Information Bottleneck for Domain Adaptive Semantic Segmentation
Yawei Luo, Ping Liu, Tao Guan, Junqing Yu, Yi Yang
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 SA-layer (significance-aware layer) to standard IB (ablation) | IBAN ResNet-101 · GTA5 -> Cityscapes (ablation, ResNet-101) | mIoU: 40.7 -> 42.2 | Improved |
| update beta adaptively via dual gradient descent (ablation) | SIBAN ResNet-101 · GTA5 -> Cityscapes (ablation, ResNet-101) | mIoU: 42.2 -> 42.6 | Improved |
| add extra discriminator on significance tensors (ablation) | SIBAN ResNet-101 · GTA5 -> Cityscapes (ablation, ResNet-101) | mIoU: 42.6 -> 43.2 | Improved |
| add extra discriminator on segmentation maps (ablation) | SIBAN ResNet-101 · GTA5 -> Cityscapes (ablation, ResNet-101) | mIoU: 43.2 -> 45.5 | Improved |
| add information bottleneck to the feature-space adversarial baseline | Baseline ResNet-101 · GTA5 -> Cityscapes | mIoU: 39.3 -> 40.7 | Improved |
| add standard information bottleneck to baseline | Baseline VGG-16 · GTA5 -> Cityscapes | mIoU: 31.7 -> 32.1 | Improved |
| replace standard IB with significance-aware IB | IBAN VGG-16 · GTA5 -> Cityscapes | mIoU: 32.1 -> 34.2 | Improved |
| replace standard IB with significance-aware IB | IBAN ResNet-101 · GTA5 -> Cityscapes | mIoU: 40.7 -> 42.6 | Improved |
| replace standard IB with significance-aware IB | IBAN VGG-16 · SYNTHIA -> Cityscapes | mIoU: 35.4 -> 37.2 | Improved |
| replace standard IB with significance-aware IB | IBAN ResNet-101 · SYNTHIA -> Cityscapes | mIoU: 43.2 -> 46.3 | Improved |
| add SIBAN to source-only model | Source only VGG-16 · GTA5 -> Cityscapes | mIoU: 17.9 -> 34.2 | Improved |
| add SIBAN to source-only model | Source only ResNet-101 · GTA5 -> Cityscapes | mIoU: 36.6 -> 42.6 | Improved |
Closest papers by meaning
- Category-Level Adversarial Adaptation for Semantic Segmentation using Purified Features (TPAMI 2021 · similarity 0.91)
- An Adversarial Perturbation Oriented Domain Adaptation Approach for Semantic Segmentation (AAAI 2020 · similarity 0.89)
- Learning to Adapt Structured Output Space for Semantic Segmentation (CVPR 2018 · similarity 0.88)
- Taking a Closer Look at Domain Shift: Category-Level Adversaries for Semantics Consistent Domain Adaptation (CVPR 2019 · similarity 0.88)
- Classes Matter: A Fine-grained Adversarial Approach to Cross-domain Semantic Segmentation (ECCV 2020 · similarity 0.87)
- Joint Adversarial Learning for Domain Adaptation in Semantic Segmentation (AAAI 2020 · 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.
