Research map / Image generation / GANs
Seeing What a GAN Cannot Generate
David Bau, Jun-Yan Zhu, Jonas Wulff, William Peebles, Hendrik Strobelt, Bolei Zhou, Antonio Torralba
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| use StyleGAN instead of Progressive GAN for bedroom generation | LSUN bedroom GAN · LSUN bedrooms | Frechet Segmentation Distance: 85.2 -> 22.6 | Improved |
| use Progressive GAN instead of WGAN-GP for bedroom generation | LSUN bedroom GAN · LSUN bedrooms | Frechet Segmentation Distance: 428.4 -> 85.2 | Improved |
| add layer-wise relaxed inversion (invert G_f and optimize perturbations of r) on top of the layer-wise encoder (ablation) | Progressive GAN inversion · LSUN churches generated images | reconstruction accuracy: 95.5% -> above 99% | Improved |
| add optimization of z after layer-wise encoder initialization (no relaxation to G_f) (ablation) | Progressive GAN inversion · LSUN churches generated images | recovery of z | No clear effect |
| use layer-wise training of the inversion network instead of end-to-end direct training (ablation) | inversion network E · LSUN churches generated images | reconstruction quality | Improved |
| initialize z optimization with encoder output E(x) instead of random | direct z optimization on 15-layer Progressive GAN · LSUN churches generated images | reconstruction quality | No clear effect |
Closest papers by meaning
- DynGAN: Solving Mode Collapse in GANs With Dynamic Clustering (TPAMI 2024 · similarity 0.90)
- GAN Dissection: Visualizing and Understanding Generative Adversarial Networks (ICLR 2019 · similarity 0.90)
- Escaping from Collapsing Modes in a Constrained Space (ECCV 2018 · similarity 0.89)
- No Modes Left Behind: Capturing the Data Distribution Effectively Using GANs (AAAI 2018 · similarity 0.89)
- A Classification-Based Study of Covariate Shift in GAN Distributions (ICML 2018 · similarity 0.88)
- The GAN is dead; long live the GAN! A Modern GAN Baseline (NeurIPS 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.
