Research map / Label-efficient and robust learning / Domain adaptation and generalization
PixMatch: Unsupervised Domain Adaptation via Pixelwise Consistency Training
Check 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 target-domain pixelwise consistency training with heavy augmentations to source-only training | DeepLab-v2, ResNet-101 · GTA5-to-Cityscapes | mIoU: 36.9 -> 48.3 | Helped |
| Add MaxSquare loss to augmentation-based PixMatch (ablation) | PixMatch (Augmentations) · GTA5-to-Cityscapes | mIoU: 49.9 -> 50.7 | Helped |
| Use CutMix instead of heavy augmentations as the consistency perturbation (ablation) | PixMatch · GTA5-to-Cityscapes | mIoU: 48.3 -> 45.4 | Hurt |
| Use Fourier perturbation instead of heavy augmentations (ablation) | PixMatch · GTA5-to-Cityscapes | mIoU: 48.3 -> 44.0 | Hurt |
| Use style-transfer perturbation instead of heavy augmentations (ablation) | PixMatch · GTA5-to-Cityscapes | mIoU: 48.3 -> 39.2 | Hurt |
| Use CutMix instead of heavy augmentations as the consistency perturbation (ablation) | PixMatch · SYNTHIA-to-Cityscapes | mIoU-13: 50.8 -> 51.2 | Helped |
| Combine Fourier and CutMix perturbations (ablation) | PixMatch · SYNTHIA-to-Cityscapes | mIoU-13: 51.1 -> 54.5 | Helped |
| Add a third augmentation perturbation to Fourier + CutMix (ablation) | PixMatch · SYNTHIA-to-Cityscapes | mIoU-13: 54.5 -> 52.1 | Hurt |
| Increase consistency-loss weight from λT 0.00 to 0.15 (ablation) | PixMatch (Augmentations) · GTA5-to-Cityscapes | mIoU: 36.9 -> 49.8 | Helped |
| Set consistency-loss weight to λT 0.20 rather than 0.15 (ablation) | PixMatch (Augmentations) · GTA5-to-Cityscapes | mIoU: 49.8 -> 48.0 | Hurt |
| Add confidence thresholding to pseudolabel consistency loss (ablation) | PixMatch (Augmentations) · GTA5-to-Cityscapes | mIoU: 49.9 -> 47.0 | Hurt |
| Use MaxSquare loss weight λMSL 0.05 rather than 0.00 (ablation) | PixMatch (Augmentations) · GTA5-to-Cityscapes | mIoU: 49.9 -> 50.7 | Helped |
Closest papers by meaning
- Consistency Regularization with High-dimensional Non-adversarial Source-guided Perturbation for Unsupervised Domain Adaptation in Segmentation (AAAI 2021 · similarity 0.91)
- Multi-Target Domain Adaptation With Collaborative Consistency Learning (CVPR 2021 · similarity 0.91)
- CrDoCo: Pixel-Level Domain Transfer With Cross-Domain Consistency (CVPR 2019 · similarity 0.90)
- Unsupervised Model Adaptation for Continual Semantic Segmentation (AAAI 2021 · similarity 0.90)
- Universal Domain Adaptation for Semantic Segmentation (CVPR 2025 · similarity 0.89)
- Open-Set Domain Adaptation for Semantic Segmentation (CVPR 2024 · similarity 0.88)
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.