Research map / Image generation / Editing and style transfer
Conditional Image-to-Image Translation
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 conditional dual-GAN instead of DualGAN, DualGAN-c, or GAN-c | Conditional image-to-image translation · CelebA men-to-women and women-to-men face translation | qualitative preservation of source features and use of conditional features | Helped |
| Use different target-domain conditional images for a fixed source image | cd-GAN · Women-to-men faces and edges-to-handbags | conditional control and diversity | Helped |
| Replace the skip connections from original domain-specific features with connections from reconstructed features (ablation) | cd-GAN-rec · Edges-to-handbags | image reconstruction quality | Hurt |
| Remove domain-specific and domain-independent feature reconstruction losses (ablation) | cd-GAN-nof · Edges-to-handbags | color schemes | Hurt |
| Remove the domain-specific feature reconstruction loss (ablation) | cd-GAN-nos · Edges-to-handbags | color schemes | Hurt |
| Remove the domain-independent feature reconstruction loss (ablation) | cd-GAN-noi · Edges-to-handbags | similarity to cd-GAN results | No clear effect |
| Set domain-independent features to zero and generate from domain-specific features only (ablation) | cd-GAN · Feature visualization | style similarity to the conditional image | Helped |
| Set domain-specific features to zero and generate from domain-independent features only (ablation) | cd-GAN · Feature visualization | preservation of source-image shape | Helped |
Closest papers by meaning
- Image-To-Image Translation With Conditional Adversarial Networks (CVPR 2017 · similarity 0.90)
- TraVeLGAN: Image-To-Image Translation by Transformation Vector Learning (CVPR 2019 · similarity 0.90)
- DINO: A Conditional Energy-Based GAN for Domain Translation (ICLR 2021 · similarity 0.89)
- Unsupervised Image-to-Image Translation With Generative Prior (CVPR 2022 · similarity 0.89)
- StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation (CVPR 2018 · similarity 0.89)
- Rethinking the Paradigm of Content Constraints in Unpaired Image-to-Image Translation (AAAI 2024 · 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.