Research map / Label-efficient and robust learning / Domain adaptation and generalization
Fully Convolutional Adaptation Networks 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 Adaptive Batch Normalization (ablation) | FCN for GTA5-to-Cityscapes segmentation · Cityscapes val | mIoU: 29.15 -> 35.51 | Helped |
| Add image-level adversarial domain adaptation (ablation) | FCN + ABN · Cityscapes val | mIoU: 5.78% | Helped |
| Classify image regions rather than whole images in the domain discriminator (ablation) | FCN + ABN + ADA · Cityscapes val | mIoU: 1.88% | Helped |
| Add ASPP to enlarge the discriminator field of view (ablation) | FCN + ABN + ADA + Conv · Cityscapes val | mIoU: 1.64% | Helped |
| Add appearance-level adaptation with AAN (ablation) | RAN with ABN, ADA, Conv, and ASPP · Cityscapes val | mIoU: 1.79% | Helped |
| Adapt source images with AAN (ablation) | FCN · Cityscapes val | mIoU: 29.15 -> 31.71 | Helped |
| Adapt target images with AAN (ablation) | FCN · Cityscapes val | mIoU: 29.15 -> 34.68 | Helped |
| Adapt source images with AAN (ablation) | RAN · Cityscapes val | mIoU: 44.81 -> 46.21 | Helped |
| Adapt target images with AAN (ablation) | RAN · Cityscapes val | mIoU: 44.81 -> 45.03 | No clear effect |
| Adapt both source and target images with AAN (ablation) | RAN · Cityscapes val | mIoU: 44.81 -> 45.59 | Helped |
| Late-fuse score maps from the four AAN settings (ablation) | FCAN · Cityscapes val | mIoU: 46.21 -> 46.60 | Helped |
| Use multi-scale inference (ablation) | FCAN · Cityscapes val | mIoU: 46.60 -> 47.75 | Helped |
Closest papers by meaning
- Learning From Synthetic Data: Addressing Domain Shift for Semantic Segmentation (CVPR 2018 · similarity 0.89)
- ROAD: Reality Oriented Adaptation for Semantic Segmentation of Urban Scenes (CVPR 2018 · similarity 0.89)
- Multi-source Domain Adaptation for Semantic Segmentation (NeurIPS 2019 · similarity 0.89)
- Conditional Generative Adversarial Network for Structured Domain Adaptation (CVPR 2018 · similarity 0.88)
- Knowledge Adaptation for Efficient Semantic Segmentation (CVPR 2019 · similarity 0.88)
- Learning to Adapt Structured Output Space for Semantic Segmentation (CVPR 2018 · 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.