Research map / Image generation / GANs
Learning Modulated Transformation in GANs
Read the paperCheck 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 MTM to StyleGAN2 low-resolution layers | StyleGAN2 · ImageNet-128 | FID: 21.14 -> 19.16 | Helped |
| add MTM to StyleGAN3 | StyleGAN3 · TaiChi-256 | FID: 21.36 -> 13.60 | Helped |
| add MTM to StyleGAN2 | StyleGAN2 · LSUN Church-256 | FID: 4.04 -> 2.32 | Helped |
| add MTM to StyleGAN2 | StyleGAN2 · LSUN Cat-256 | FID: 6.87 -> 5.92 | Helped |
| add MTM to EG3D | EG3D · FFHQ-256 | FID: 4.32 -> 4.07 | Helped |
| add MTM to StyleSV | StyleSV · SkyTimelapse | FVD16: 49.0 -> 42.3 | Helped |
| add MTM to StyleSV | StyleSV · TaiChi-256 video | FVD16: 97.4 -> 89.5 | Helped |
| add MTM to StyleSV | StyleSV · YouTube Driving | FID: 19.2 -> 10.3 | Helped |
| also apply MTM to Mid-resolution layers (ablation) | StyleGAN2 w/ MTM (Low) · ImageNet-128 | FID: 19.16 -> 19.35 | Hurt |
| also apply MTM to High-resolution layers (ablation) | StyleGAN2 w/ MTM (Low+Mid) · ImageNet-128 | FID: 19.35 -> 20.02 | Hurt |
| apply MTM in more layer groups (ablation) | StyleGAN2 w/ MTM · ImageNet-128 | training time | Hurt |
| set learned offsets to zero after training (ablation) | MTM generator · TaiChi-256 images | synthesis quality | Hurt |
Closest papers by meaning
- A Geometric Analysis of Deep Generative Image Models and Its Applications (ICLR 2021 · similarity 0.88)
- Generative Adversarial Transformers (ICML 2021 · similarity 0.88)
- Navigating the GAN Parameter Space for Semantic Image Editing (CVPR 2021 · similarity 0.88)
- Diffusion-driven GAN Inversion for Multi-Modal Face Image Generation (CVPR 2024 · similarity 0.87)
- WeditGAN: Few-Shot Image Generation via Latent Space Relocation (AAAI 2024 · similarity 0.87)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · similarity 0.87)
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.