Research map / Image generation / Autoregressive and latent generation
Conditional Adversarial Generative Flow for Controllable Image Synthesis
Rui Liu, Yu Liu, Xinyu Gong, Xiaogang Wang, Hongsheng Li
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 CAGlow instead of CGlow for conditional generation | Glow · MNIST | accuracy: 98.89% -> 99.55% | Improved |
| Use CAGlow instead of CGlow for conditional generation | Glow · CelebA | accuracy: 87.43% -> 95.16% | Improved |
| Use CAGlow instead of Glow | Glow · MNIST | FID: 25.78 -> 26.34 | Hurt |
| Use CAGlow instead of CGlow | Glow · MNIST | FID: 29.64 -> 26.34 | Improved |
| Use CAGlow instead of Glow | Glow · CelebA | FID: 103.67 -> 104.91 | Hurt |
| Use CAGlow instead of CGlow | Glow · CelebA | FID: 126.52 -> 104.91 | Improved |
| Use CAGlow instead of CGlow | conditional generation · CelebA | attribute accuracy: 87.36% -> 93.75% | Improved |
| Use CAGlow instead of CGlow | conditional generation · CelebA | AMP variance: 0.0245 -> 0.0016 | Improved |
| Use CAGlow instead of the strongest baseline, Glow + pre-store | cumulative attribute manipulation · CelebA, 1 to 4 manipulations | absolute AMP difference: 0.002245 -> 0.000774; 0.007213 -> 0.002608; 0.014352 -> 0.005825; 0.023767 -> 0.009939 | Improved |
Closest papers by meaning
- Continuous Conditional Generative Adversarial Networks: Novel Empirical Losses and Label Input Mechanisms (TPAMI 2022 · similarity 0.89)
- Dual Projection Generative Adversarial Networks for Conditional Image Generation (ICCV 2021 · similarity 0.89)
- CcGAN: Continuous Conditional Generative Adversarial Networks for Image Generation (ICLR 2021 · similarity 0.89)
- Robust Conditional Generative Adversarial Networks (ICLR 2019 · similarity 0.88)
- Attack Deterministic Conditional Image Generative Models for Diverse and Controllable Generation (AAAI 2024 · similarity 0.87)
- Turning Waste into Wealth: Leveraging Low-Quality Samples for Enhancing Continuous Conditional Generative Adversarial Networks (AAAI 2024 · 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.
