Research map / Image generation / GANs
Intriguing Property and Counterfactual Explanation of GAN for Remote Sensing Image Generation
Xingzhe Su, Wenwen Qiang, Jie Hu, Changwen Zheng, Fengge Wu, Fuchun Sun
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 uniformity and entropy regularization on generator features | StyleGAN2+ADA · NWPU-RESISC45 (full) | FID: 11.97 -> -2.23 (improvement 2.23) | Improved |
| add the regularization | BigGAN+ADA · NWPU-RESISC45 | FID: 30.91 -> -2.09 (improvement 2.09) | Improved |
| add the regularization on top of LeCam | StyleGAN2+ADA+LeCam · PatternNet | FID: 31.32 -> -2.11 (improvement 2.11) | Improved |
| add the regularization on top of APA augmentation | StyleGAN2+APA · AID | FID: 32.51 -> -3.22 (improvement 3.22) | Improved |
| add the regularization to other modalities (SAR, thermal, UAV) | StyleGAN2-ADA · HIT-UAV | FID: 83.77 -> 81.57 | Improved |
| add the regularization on natural images | StyleGAN2-ADA · FFHQ256 (2k images) | FID: 26.41 -> 24.20 | Improved |
| add uniformity regularization (ablation) | StyleGAN2+ADA · NWPU-RESISC45 (300 samples) | FID: 20.63 -> 17.86 | Improved |
| add entropy regularization (ablation) | StyleGAN2+ADA · PatternNet (full) | FID: 33.53 -> 32.87 | Improved |
| add both uniformity and entropy regularization (ablation) | StyleGAN2+ADA · PatternNet (18 classes) | FID: 28.88 -> 25.22 | Improved |
| add entropy regularization to uniformity-only variant (ablation) | StyleGAN2+ADA · PatternNet (full) | FID: 31.48 -> 32.55 | Hurt |
| lower temperature gamma from 7 to 2 (ablation) | StyleGAN2+ADA · NWPU-RESISC45 (20 classes) | FID: 24.96 -> 15.05 | Improved |
| use both generator and discriminator features instead of generator only (ablation) | StyleGAN2+ADA · NWPU-RESISC45 (full) | FID: 14.78 -> 10.74 | Improved |
Closest papers by meaning
- Remote Sensing Image Generation via Object Text Decoupling (TPAMI 2025 · similarity 0.86)
- On Leveraging Pretrained GANs for Generation with Limited Data (ICML 2020 · similarity 0.85)
- MetaEarth: A Generative Foundation Model for Global-Scale Remote Sensing Image Generation (TPAMI 2024 · similarity 0.85)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · similarity 0.84)
- RG-GAN: Dynamic Regenerative Pruning for Data-Efficient Generative Adversarial Networks (AAAI 2024 · similarity 0.84)
- Feature Spectrum Learning for Remote Sensing Change Detection (CVPR 2025 · similarity 0.84)
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.
