Research map / Image generation / Autoregressive and latent generation
Image Synthesis Under Limited Data: A Survey and Taxonomy
Mengping Yang, Zhe Wang
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 ADA augmentation to StyleGAN2 | StyleGAN2 · FFHQ 1K images | FID: 100.16 -> 21.29 | Improved |
| add InsGen/FakeCLR regularization on top of ADA | ADA · FFHQ 1K images | FID: 21.29 -> 15.92 | Improved |
| use off-the-shelf pretrained features (ProjectedGAN) instead of augmentation | StyleGAN-ADA · AFHQ Cat | FID: 23.34 -> 2.16 | Improved |
| add Vision-Aided discriminator to ADA | StyleGAN-ADA · AFHQ Cat | FID: 23.34 -> 2.69 | Improved |
| add ADA augmentation to StyleGAN2 | StyleGAN2 · CIFAR-10 20% data | FID: 19.37 -> 7.40 | Improved |
| add FakeCLR regularization | StyleGAN2 · Animal Faces Cat | FID: 69.84 -> 26.34 | Improved |
| adapt a pretrained generator (TransferGAN) instead of training from scratch | StyleGAN2 · 100-shot Obama | FID: 80.45 -> 48.73 | Improved |
| replace fine-tuning (TransferGAN) with regularization (CDC) | few-shot adaptation · FFHQ to Babies | FID: 104.79 -> 74.39 | Improved |
| replace fine-tuning (TransferGAN) with regularization (CDC) | few-shot adaptation · FFHQ to AFHQ-Cat (distant domain) | FID: 64.68 -> 176.21 | Hurt |
| use kernel modulation (RICK) instead of fine-tuning | few-shot adaptation · FFHQ to AFHQ-Cat (distant domain) | FID: 64.68 -> 53.27 | Improved |
| use regularization (CDC) instead of fine-tuning | few-shot adaptation · FFHQ to Otto's Paintings | LPIPS: 0.51 -> 0.63 | Improved |
Closest papers by meaning
- Any-Resolution Training for High-Resolution Image Synthesis (ECCV 2022 · similarity 0.89)
- Plug & Play Generative Networks: Conditional Iterative Generation of Images in Latent Space (CVPR 2017 · similarity 0.89)
- Diffusion Models in Low-Level Vision: A Survey (TPAMI 2025 · similarity 0.89)
- Novel View Synthesis with Pixel-Space Diffusion Models (CVPR 2025 · similarity 0.89)
- Boosting Generative Image Modeling via Joint Image-Feature Synthesis (NeurIPS 2025 · similarity 0.88)
- ScaleCrafter: Tuning-free Higher-Resolution Visual Generation with Diffusion Models (ICLR 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.
