Research map / Image generation / GANs
A U-Net Based Discriminator for Generative Adversarial Networks
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 |
|---|---|---|---|
| Replace the image-classification discriminator with a U-Net that gives global and per-pixel feedback (ablation) | BigGAN · FFHQ | FID: 12.42 -> 10.86 | Helped |
| Replace the image-classification discriminator with a U-Net that gives global and per-pixel feedback (ablation) | BigGAN · COCO-Animals | FID: 16.55 -> 15.86 | Helped |
| Add CutMix augmentation (ablation) | U-Net discriminator · FFHQ | FID: 10.86 -> 10.30 | Helped |
| Add CutMix augmentation (ablation) | U-Net discriminator · COCO-Animals | FID: 15.86 -> 14.95 | Helped |
| Add per-pixel consistency regularization on CutMix images (ablation) | U-Net discriminator + CutMix · FFHQ | FID: 10.30 -> 7.63 | Helped |
| Add per-pixel consistency regularization on CutMix images (ablation) | U-Net discriminator + CutMix · COCO-Animals | FID: 14.95 -> 13.87 | Helped |
| Use U-Net GAN instead of BigGAN | BigGAN · FFHQ | IS: 3.97 -> 4.46 | Helped |
| Use U-Net GAN instead of BigGAN | BigGAN · COCO-Animals | IS: 11.77 -> 12.29 | Helped |
| Use U-Net GAN instead of BigGAN | BigGAN · FFHQ | FID: 11.48 -> 7.48 | Helped |
| Use U-Net GAN instead of BigGAN | BigGAN · COCO-Animals | FID: 16.37 -> 13.73 | Helped |
| Use U-Net GAN instead of BigGAN | BigGAN · CelebA (128×128, TensorFlow) | FID: 4.54 -> 2.95 | Helped |
| Remove the self-attention layer (ablation) | U-Net GAN encoder and decoder · Experiments described in the paper | performance | No clear effect |
Closest papers by meaning
- Large Scale GAN Training for High Fidelity Natural Image Synthesis (ICLR 2019 · similarity 0.89)
- Training Generative Adversarial Networks with Limited Data (NeurIPS 2020 · similarity 0.88)
- SGAN: An Alternative Training of Generative Adversarial Networks (CVPR 2018 · similarity 0.88)
- Conditional Image Synthesis with Auxiliary Classifier GANs (ICML 2017 · similarity 0.88)
- Progressive Growing of GANs for Improved Quality, Stability, and Variation (ICLR 2018 · similarity 0.88)
- Matching Adversarial Networks (CVPR 2018 · 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.