Research map / Image generation / GANs
High-Resolution Image Synthesis and Semantic Manipulation With Conditional GANs
Read the paperCheck 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 |
|---|---|---|---|
| use the proposed high-resolution GAN framework instead of pix2pix or CRN | conditional image synthesis from semantic maps · Cityscapes | Mean IoU: pix2pix 0.3948; CRN 0.3483; Ours 0.6389 | Helped |
| replace U-Net with the coarse-to-fine generator (ablation) | pix2pixHD framework · Cityscapes | Mean IoU: 0.3905 -> 0.6389 | Helped |
| replace CRN generator with the coarse-to-fine generator (ablation) | pix2pixHD framework · Cityscapes | Mean IoU: 0.3994 -> 0.6389 | Helped |
| add multi-scale discriminators instead of using a single discriminator (ablation) | pix2pixHD framework · Cityscapes | Mean IoU: 0.5775 -> 0.6389 | Helped |
| add multi-scale discriminators instead of using a single discriminator (ablation) | pix2pixHD framework · Cityscapes | human preference: 69.2% | Helped |
| add discriminator feature matching loss to GAN loss (ablation) | pix2pixHD framework · Cityscapes | human preference: 68.55% | Helped |
| add VGG perceptual loss to GAN plus feature matching loss (ablation) | pix2pixHD framework · Cityscapes | human preference: 58.90% | Helped |
| omit VGG perceptual loss (ablation) | pix2pixHD framework · Cityscapes | visual quality: 94.6% vs pix2pix; 85.2% vs CRN | No clear effect |
| add instance boundary maps to semantic labels (ablation) | pix2pixHD framework · Cityscapes | human preference for car-region realism: 64.34% | Helped |
Closest papers by meaning
- Linear Semantics in Generative Adversarial Networks (CVPR 2021 · similarity 0.90)
- You Only Need Adversarial Supervision for Semantic Image Synthesis (ICLR 2021 · similarity 0.90)
- Conditional Image Synthesis with Auxiliary Classifier GANs (ICML 2017 · similarity 0.89)
- Learning to Predict Layout-to-image Conditional Convolutions for Semantic Image Synthesis (NeurIPS 2019 · similarity 0.89)
- Image-To-Image Translation With Conditional Adversarial Networks (CVPR 2017 · similarity 0.89)
- Unlocking Pre-trained Image Backbones for Semantic Image Synthesis (CVPR 2024 · similarity 0.89)
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.