Research map / Label-efficient and robust learning / Semi-supervised learning
KE-GAN: Knowledge Embedded Generative Adversarial Networks for Semi-Supervised Scene Parsing
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 |
|---|---|---|---|
| Use KE-GAN with semi-supervised training instead of the strongest listed semi-supervised baseline | KE-GAN · SiftFlow | mIoU: 23.2 -> 27.2 | Helped |
| Use KE-GAN with semi-supervised training instead of the strongest listed semi-supervised baseline | KE-GAN · CamVid | mIoU: 58.2 -> 61.9 | Helped |
| Use KE-GAN with semi-supervised training instead of KE-GAN fully-supervised training | KE-GAN · CamVid | mIoU: 60.2 -> 61.9 | Helped |
| Add adversarial training losses to the pyramid baseline (ablation) | KE-GAN · Cityscapes, 1/8, 1/4, 1/2, and Full labeled | mIoU: 62.6 -> 66.1; 65.5 -> 69.2; 69.9 -> 71.5; 72.6 -> 73.6 | Helped |
| Add the knowledge relation loss to the pyramid baseline (ablation) | KE-GAN · Cityscapes, 1/8, 1/4, 1/2, and Full labeled | mIoU: 62.6 -> 63.7; 65.5 -> 66.3; 69.9 -> 70.3; 72.6 -> 72.9 | Helped |
| Add pyramid pooling to the baseline (ablation) | Cityscapes scene parsing · Cityscapes | mIoU | Helped |
| Add both adversarial training losses and knowledge relation loss to the pyramid baseline (ablation) | KE-GAN · Cityscapes, 1/8, 1/4, 1/2, and Full labeled | mIoU: 62.6 -> 66.9; 65.5 -> 70.6; 69.9 -> 72.2; 72.6 -> 75.3 | Helped |
| Add adversarial training losses rather than only the knowledge relation loss (ablation) | KE-GAN · Cityscapes | mIoU: 63.7 -> 66.1; 66.3 -> 69.2; 70.3 -> 71.5; 72.9 -> 73.6 | Helped |
Closest papers by meaning
- Generating Triples With Adversarial Networks for Scene Graph Construction (AAAI 2018 · similarity 0.88)
- Local Class-Specific and Global Image-Level Generative Adversarial Networks for Semantic-Guided Scene Generation (CVPR 2020 · similarity 0.88)
- Linear Semantics in Generative Adversarial Networks (CVPR 2021 · similarity 0.88)
- Region-Based Semantic Factorization in GANs (ICML 2022 · similarity 0.87)
- Semantic Pyramid for Image Generation (CVPR 2020 · similarity 0.87)
- Deep Structured Scene Parsing by Learning With Image Descriptions (CVPR 2016 · 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.