Research map / Image generation / GANs
Transformation GAN for Unsupervised Image Synthesis and Representation Learning
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 feature-transformation matching (ablation) | TrGAN · CIFAR-10 | FID: 15.03 -> 13.41 | Helped |
| add intermediate feature matching (ablation) | TrGAN · CIFAR-10 | FID: 14.57 -> 13.41 | Helped |
| add intermediate feature matching and feature-transformation matching (ablation) | TrGAN · CIFAR-10 | FID: 15.67 -> 13.41 | Helped |
| replace rotation prediction with projective-transformation prediction and add feature matching regularizers | unsupervised GAN image synthesis · CIFAR-10 | FID: 15.65 -> 13.41 | Helped |
| use TrGAN rather than SS-GAN | unsupervised GAN image synthesis · ImageNet | FID: 43.87 -> 39.14 | Helped |
| use TrGAN rather than SS-GAN | unsupervised GAN image synthesis · LSUN-BEDROOM | FID: 13.30 -> 11.74 | Helped |
| use TrGAN rather than SS-GAN | unsupervised GAN image synthesis · CELEBA-HQ | FID: 24.36 -> 23.17 | Helped |
| use TrGAN rather than Cond-GAN | unsupervised GAN image synthesis · CIFAR-10 | FID: 15.53 -> 13.41 | Helped |
| use TrGAN rather than Cond-GAN | unsupervised GAN image synthesis · ImageNet | FID: 42.91 -> 39.14 | Helped |
| use TrGAN rather than AET-only | unsupervised representation learning · CIFAR-10 Block0 | top-1 classification accuracy: 78.16 -> 78.74 | Helped |
| use TrGAN rather than AET-only | unsupervised representation learning · CIFAR-10 Block1 | top-1 classification accuracy: 81.92 -> 82.73 | Helped |
| use TrGAN rather than AET-only | unsupervised representation learning · CIFAR-10 Block2 | top-1 classification accuracy: 87.96 -> 87.12 | Hurt |
Closest papers by meaning
- Self-Supervised GANs via Auxiliary Rotation Loss (CVPR 2019 · similarity 0.91)
- Conditional Image Synthesis with Auxiliary Classifier GANs (ICML 2017 · similarity 0.90)
- TraVeLGAN: Image-To-Image Translation by Transformation Vector Learning (CVPR 2019 · similarity 0.90)
- Large Scale Adversarial Representation Learning (NeurIPS 2019 · similarity 0.90)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · similarity 0.88)
- Unlocking Pre-trained Image Backbones for Semantic Image Synthesis (CVPR 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.