Research map / Image generation / GANs
ReMix: Towards Image-to-Image Translation With Limited Data
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 |
|---|---|---|---|
| Add ReMix feature interpolation and relative content loss | StarGAN v2 · AFHQ latent-guided translation, 10% data | FID: 46.02 -> 21.82 | Helped |
| Add ReMix feature interpolation and relative content loss | StarGAN v2 · AFHQ latent-guided translation, 10% data | LPIPS: 0.431 -> 0.471 | Helped |
| Add ReMix feature interpolation and relative content loss | StarGAN v2 · AFHQ reference-guided translation, 10% data | FID: 38.42 -> 22.92 | Helped |
| Add ReMix feature interpolation and relative content loss | LightCNN + StyleGAN2 · CASIA NIR-VIS 2.0, first fold | Rank-1 accuracy (%): 93.13 -> 98.18 | Helped |
| Add ReMix input interpolation and relative perceptual loss | SPADE · Cityscapes, 10% training data | mIoU: 48.3 -> 62.1 | Helped |
| Add ReMix input interpolation and relative perceptual loss | SPADE · Cityscapes, 10% training data | FID: 85.9 -> 68.0 | Helped |
| Apply WM augmentation | StarGAN v2 · AFHQ reference-guided translation, 10% data | FID: 38.42 -> 45.88 | Hurt |
| Apply WM augmentation | LightCNN + StyleGAN2 · CASIA NIR-VIS 2.0, first fold | Rank-1 accuracy (%): 93.13 -> 91.32 | Hurt |
| Apply mixup augmentation | SPADE · Cityscapes, 10% training data | mIoU: 48.3 -> 59.7 | Helped |
| Use ReMix with limited training data | SPADE · Cityscapes, 10% versus 100% training data | mIoU: 70.3 -> 62.1 | Mixed |
Closest papers by meaning
- Rethinking the Paradigm of Content Constraints in Unpaired Image-to-Image Translation (AAAI 2024 · similarity 0.90)
- DeepI2I: Enabling Deep Hierarchical Image-to-Image Translation by Transferring from GANs (NeurIPS 2020 · similarity 0.90)
- TraVeLGAN: Image-To-Image Translation by Transformation Vector Learning (CVPR 2019 · similarity 0.90)
- E$^2$GAN: Efficient Training of Efficient GANs for Image-to-Image Translation (ICML 2024 · similarity 0.89)
- Unsupervised Image-to-Image Translation With Generative Prior (CVPR 2022 · similarity 0.89)
- DA-GAN: Instance-Level Image Translation by Deep Attention Generative Adversarial Networks (CVPR 2018 · 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.