Research map / Image generation / GANs
Stochastic Conditional Generative Networks with Basis Decomposition
Ze Wang, Xiuyuan Cheng, Guillermo Sapiro, Qiang Qiu
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 |
|---|---|---|---|
| Replace Pix2Pix convolution layers with stochastic basis-generator layers | Pix2Pix · Labels → Facade | LPIPS: 0.0003 ± 0.0000 -> 0.2648 ± 0.004 | Improved |
| Replace Pix2Pix convolution layers with stochastic basis-generator layers | Pix2Pix · Map → Satellite | FID: 168.99 ± 2.58 -> 35.54 ± 2.19 | Improved |
| Use BasisGAN rather than MUNIT | Pix2Pix · Edge → Shoe | FID: 62.57 ± 0.917 -> 64.17 ± 1.14 | Hurt |
| Generate 20 samples per input with BasisGAN | BasisGAN · Labels → Facade | FID: 88.7 ± 1.28 -> 24.14 ± 0.76 | Improved |
| Increase basis-generator dimensions from 64 + 64 to 16 + 16 (ablation) | BasisGAN · Map → Satellite | LPIPS: 0.2417 -> 0.2242 | Hurt |
| Increase basis-generator dimensions from 64 + 64 to 32 + 32 (ablation) | BasisGAN · Map → Satellite | LPIPS: 0.2417 -> 0.2388 | Hurt |
| Increase basis-generator dimensions from 64 + 64 to 128 + 128 (ablation) | BasisGAN · Map → Satellite | LPIPS: 0.2417 -> 0.2448 | Improved |
| Increase basis-generator dimensions from 64 + 64 to 256 + 256 (ablation) | BasisGAN · Map → Satellite | LPIPS: 0.2417 -> 0.2452 | Improved |
| Increase basis-generator dimensions from 64 + 64 to 512 + 512 (ablation) | BasisGAN · Map → Satellite | LPIPS: 0.2417 -> 0.2433 | Improved |
| Increase basis count from K=7 to K=16 (ablation) | BasisGAN · Map → Satellite | LPIPS: 0.2417 -> 0.2409 | Hurt |
| Increase basis count from K=7 to K=32 (ablation) | BasisGAN · Map → Satellite | LPIPS: 0.2417 -> 0.2382 | Hurt |
| Increase basis count from K=7 to K=64 (ablation) | BasisGAN · Map → Satellite | LPIPS: 0.2417 -> 0.2288 | Hurt |
Closest papers by meaning
- Diversity-Sensitive Conditional Generative Adversarial Networks (ICLR 2019 · similarity 0.90)
- Semi-supervised FusedGAN for Conditional Image Generation (ECCV 2018 · similarity 0.89)
- Plug & Play Generative Networks: Conditional Iterative Generation of Images in Latent Space (CVPR 2017 · similarity 0.89)
- Attack Deterministic Conditional Image Generative Models for Diverse and Controllable Generation (AAAI 2024 · similarity 0.87)
- Unlocking Pre-trained Image Backbones for Semantic Image Synthesis (CVPR 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.
