Research map / Label-efficient and robust learning / Domain adaptation and generalization
A Curriculum Domain Adaptation Approach to the Semantic Segmentation of Urban Scenes
Yang Zhang, Philip David, Hassan Foroosh, Boqing Gong
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 curriculum domain adaptation with image and landmark-superpixel label distributions | FCN-8s · GTA to Cityscapes | mIoU: 22.3 -> 28.9 | Improved |
| add curriculum domain adaptation with image and landmark-superpixel label distributions | FCN-8s · SYNTHIA to Cityscapes | mIoU: 22.0 -> 29.0 | Improved |
| add color constancy preprocessing to the full method (ablation) | Ours (I+SP) · GTA to Cityscapes | mIoU: 28.9 -> 31.4 | Improved |
| add color constancy preprocessing to the full method (ablation) | Ours (I+SP) · SYNTHIA to Cityscapes | mIoU: 29.0 -> 29.7 | Improved |
| add landmark superpixel distributions to image label distributions (ablation) | Ours (I) · GTA to Cityscapes | mIoU: 23.1 -> 28.9 | Improved |
| add global image label distributions to landmark superpixel distributions (ablation) | Ours (SP) · SYNTHIA to Cityscapes | mIoU: 28.1 -> 29.0 | Improved |
| add global image label distributions to landmark superpixel distributions (ablation) | Ours (SP) · GTA to Cityscapes | mIoU: 27.8 -> 28.9 | Improved |
| add color constancy to the no-adaptation baseline (ablation) | NoAdapt · GTA to Cityscapes | mIoU: 22.3 -> 26.2 | Improved |
| add color constancy to the no-adaptation baseline (ablation) | NoAdapt · SYNTHIA to Cityscapes | mIoU: 22.0 -> 22.6 | Improved |
| replace CNN superpixel features with handcrafted BOW+FV features (ablation) | Ours (CC) · GTA to Cityscapes | mIoU: 28.3 vs NoAdapt (CC) 26.2 | No clear effect |
| use landmark superpixels at 400 instead of 50 superpixels per image (ablation) | SP (CC) · GTA to Cityscapes | mIoU: 22.9 -> 27.4 | Improved |
| use label distributions of all superpixels instead of only landmarks (ablation) | Ours (SP) · GTA to Cityscapes | mIoU | No clear effect |
Closest papers by meaning
- Curriculum Domain Adaptation for Semantic Segmentation of Urban Scenes (ICCV 2017 · similarity 0.92)
- Effective Use of Synthetic Data for Urban Scene Semantic Segmentation (ECCV 2018 · similarity 0.92)
- ROAD: Reality Oriented Adaptation for Semantic Segmentation of Urban Scenes (CVPR 2018 · similarity 0.91)
- Constructing Self-Motivated Pyramid Curriculums for Cross-Domain Semantic Segmentation: A Non-Adversarial Approach (ICCV 2019 · similarity 0.91)
- Learning Content-Enhanced Mask Transformer for Domain Generalized Urban-Scene Segmentation (AAAI 2024 · similarity 0.90)
- Stronger Fewer & Superior: Harnessing Vision Foundation Models for Domain Generalized 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.
