Research map / Image generation / GANs
A Style-Based Generator Architecture 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 traditional generator with a style-based generator | Progressive GAN improved baseline · CelebA-HQ and FFHQ | FID: almost 20% | Helped |
| add a mapping network and AdaIN operations (ablation) | improved baseline generator · CelebA-HQ and FFHQ | FID | Helped |
| remove the traditional input layer and start from a learned constant (ablation) | generator with mapping network and AdaIN · CelebA-HQ and FFHQ | FID | Helped |
| add per-layer noise inputs (ablation) | style-based generator without noise inputs · CelebA-HQ and FFHQ | FID | Helped |
| add mixing regularization (ablation) | style-based generator · style-mixing test scenarios | FID | Helped |
| use a mapping network (ablation) | traditional generator · FFHQ | perceptual path length in Z | Hurt |
| use a mapping network (ablation) | traditional generator · FFHQ | FID | Helped |
| increase mapping-network depth (ablation) | style-based generator · FFHQ | perceptual path length | Helped |
| increase mapping-network depth (ablation) | style-based generator · CelebA-HQ | linear separability | Helped |
| train longer (ablation) | style-based generator · FFHQ | perceptual path length | Hurt |
| train longer (ablation) | style-based generator · FFHQ | FID | Helped |
Closest papers by meaning
- Adversarial Latent Autoencoders (CVPR 2020 · similarity 0.89)
- Towards Principled Methods for Training Generative Adversarial Networks (ICLR 2017 · similarity 0.88)
- Cluster-Guided Image Synthesis With Unconditional Models (CVPR 2022 · similarity 0.88)
- FIANCEE: Faster Inference of Adversarial Networks via Conditional Early Exits (CVPR 2023 · similarity 0.88)
- Learning Semantic-aware Normalization for Generative Adversarial Networks (NeurIPS 2020 · similarity 0.88)
- Improving Generative Adversarial Networks with Denoising Feature Matching (ICLR 2017 · 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.