Research map / Detection and segmentation / Weakly supervised detection and segmentation
Transferable Semi-Supervised 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 L-Net-generated cross-category labels instead of AE-generated labels | DeepLab-LargeFOV segmentation model · PASCAL VOC 2012 val, split-sets 1–3 | mIoU: 13.1%; 10.6%; 16.3% | Helped |
| Add P-Net adversarial refinement to L-Net-based segmentation (ablation) | DeepLab-LargeFOV segmentation model · PASCAL VOC 2012 validation, split-sets 1–4 | mIoU: +1.1; +1.3; +1.1; +0.9 | Helped |
| Pre-train P-Net on strong categories before fine-tuning the full training set (ablation) | P-Net · PASCAL VOC 2012 weak categories | mIoU: 0.5% | Helped |
| Train P-Net directly on all categories, omitting strong-category pretraining (ablation) | P-Net · PASCAL VOC 2012 split-sets 2–4 | mIoU | Hurt |
| Use L-Net transferable label generation rather than TransferNet | Cross-category semantic segmentation · PASCAL VOC 2012 val, split-set 4 | mIoU: 6.1% | Helped |
| Use cross-category transfer with only half the categories pixel-labeled rather than fully supervised training | Proposed model · PASCAL VOC 2012 | performance relative to fully-supervised baseline: 96.5% | Hurt |
| Use cross-dataset transfer from MS-COCO strong categories to VOC weak categories | Proposed model · PASCAL VOC 2012 split-set 4 | performance relative to fully-supervised baseline: 89.4% | Hurt |
Closest papers by meaning
- GuidedMix-Net: Semi-supervised Semantic Segmentation by Using Labeled Images as Reference (AAAI 2022 · similarity 0.90)
- Semi-Supervised Semantic Image Segmentation With Self-Correcting Networks (CVPR 2020 · similarity 0.90)
- Learning Transferrable Knowledge for Semantic Segmentation With Deep Convolutional Neural Network (CVPR 2016 · similarity 0.90)
- Combining Bottom-Up, Top-Down, and Smoothness Cues for Weakly Supervised Image Segmentation (CVPR 2017 · similarity 0.89)
- Semi-Supervised Semantic Segmentation With Cross-Consistency Training (CVPR 2020 · similarity 0.89)
- SemiDAViL: Semi-supervised Domain Adaptation with Vision-Language Guidance for Semantic Segmentation (CVPR 2025 · 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.