Research map / Label-efficient and robust learning / Semi-supervised learning
Semantic Segmentation With Generative Models: Semi-Supervised Learning and Strong Out-of-Domain Generalization
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 test-time latent optimization after encoder initialization (ablation) | SemanticGAN · CelebA, 150 labels | mIoU: 0.7016 -> 0.7600 | Helped |
| Add test-time latent optimization after encoder initialization (ablation) | SemanticGAN · MetFaces, 150 labels | mIoU: 0.5643 -> 0.6336 | Helped |
| Increase unlabeled CelebA data from 3K to 28K (ablation) | SemanticGAN with 30 labels · CelebA | mIoU: 0.6786 -> 0.6902 | Helped |
| Increase unlabeled MetFaces-related training data from 3K to 28K (ablation) | SemanticGAN with 30 labels · MetFaces-40 | mIoU: 0.5410 -> 0.5883 | Helped |
| Train a DeepLab model with synthetic images only instead of real labeled data only (ablation) | DeepLabV2 · CelebA | mIoU: 0.6591 -> 0.6829 | Helped |
| Train a DeepLab model with synthetic images only instead of real labeled data only (ablation) | DeepLabV2 · MetFaces | mIoU: 0.4988 -> 0.5137 | Helped |
| Mix synthetic data with real labeled data and sample without truncation (ablation) | DeepLabV2 · MetFaces | mIoU: 0.4988 -> 0.5656 | Helped |
Closest papers by meaning
- DatasetGAN: Efficient Labeled Data Factory With Minimal Human Effort (CVPR 2021 · similarity 0.90)
- Reference-Guided Pseudo-Label Generation for Medical Semantic Segmentation (AAAI 2022 · similarity 0.89)
- Transferable Semi-Supervised Semantic Segmentation (AAAI 2018 · similarity 0.89)
- SemiGDA: Generative Dual-distribution Alignment for Semi-Supervised Medical Image Segmentation (CVPR 2026 · similarity 0.89)
- Generative Medical Segmentation (AAAI 2025 · similarity 0.88)
- SemiDAViL: Semi-supervised Domain Adaptation with Vision-Language Guidance for Semantic Segmentation (CVPR 2025 · 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.