Research map / Image generation / GANs
Learning Conditional Generative Models for Phase Retrieval
Tobias Uelwer, Sebastian Konietzny, Alexander Oberstrass, Stefan Harmeling
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 |
|---|---|---|---|
| add latent optimization to the CGAN | PRCGAN-D · MNIST Fourier phase retrieval | MSE: 0.0185 -> 0.0009 | Improved |
| add latent optimization to the CGAN | PRCGAN-D · FMNIST Fourier phase retrieval | MSE: 0.0149 -> 0.0084 | Improved |
| add weight optimization to the CGAN | PRCGAN-D · CelebA Fourier phase retrieval | MSE: 0.0155 -> 0.0115 | Improved |
| latent optimization on CIFAR-10 | PRCGAN-D · CIFAR-10 Fourier phase retrieval | MSE: 0.0402 -> 0.0489 | Hurt |
| replace E2E regression with PRCGAN-D adversarial training | E2E · MNIST Fourier phase retrieval | MSE: 0.0164 -> 0.0185 | No clear effect |
| add adversarial loss (ablation) | PRCGAN-L · MNIST Fourier phase retrieval | MSE: 0.0289 -> 0.0009 | Improved |
| condition on magnitudes rather than labels (ablation) | PRCGAN-L · MNIST Fourier phase retrieval | MSE: 0.0160 -> 0.0009 | Improved |
| feed correct magnitudes to the generator (ablation) | PRCGAN-L · MNIST Fourier phase retrieval | MSE: 0.0244 -> 0.0009 | Improved |
| use LPIPS instead of MAE reconstruction loss (ablation) | PRCGAN-D · CelebA Fourier phase retrieval | MSE: 0.0155 -> 0.0143 | Improved |
| use LPIPS instead of MAE reconstruction loss (ablation) | PRCGAN-L · CelebA Fourier phase retrieval | MSE: 0.0093 -> 0.0096 | No clear effect |
| use MSE instead of MAE reconstruction loss (ablation) | PRCGAN-L · CelebA Fourier phase retrieval | MSE: 0.0093 -> 0.0125 | Hurt |
| add intermediate fully connected layers (ablation) | PRCGAN-D · CelebA Fourier phase retrieval | MSE: 0.0260 -> 0.0155 | Improved |
Closest papers by meaning
- InvNet: Encoding Geometric and Statistical Invariances in Deep Generative Models (AAAI 2020 · similarity 0.87)
- Phase Retrieval Under a Generative Prior (NeurIPS 2018 · similarity 0.87)
- pcaGAN: Improving Posterior-Sampling cGANs via Principal Component Regularization (NeurIPS 2024 · similarity 0.86)
- Unsupervised Deep Learning for Phase Retrieval via Teacher-Student Distillation (AAAI 2023 · similarity 0.86)
- prDeep: Robust Phase Retrieval with a Flexible Deep Network (ICML 2018 · similarity 0.86)
- A Regularized Conditional GAN for Posterior Sampling in Image Recovery Problems (NeurIPS 2023 · similarity 0.86)
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.
