Research map / Image generation / Editing and style transfer
Multi-mapping Image-to-Image Translation via Learning Disentanglement
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 |
|---|---|---|---|
| add translation path (random cross-domain translation) (ablation) | DMIT · Yosemite summer->winter | LPIPS: 0.109 -> 0.302 | Helped |
| add translation path (ablation) | DMIT · Yosemite summer->winter | FID: 75.90 -> 58.46 | Helped |
| add disentanglement path (ablation) | DMIT · Yosemite summer->winter | FID: 116.71 -> 58.46 | Helped |
| add disentanglement path (ablation) | DMIT · Yosemite summer->winter | LPIPS: 0.545 -> 0.302 | Hurt |
| add content adversarial loss L_GAN^c (ablation) | DMIT · Yosemite summer->winter | LPIPS: 0.268 -> 0.302 | Helped |
| add content adversarial loss L_GAN^c (ablation) | DMIT · Yosemite summer->winter | FID: 60.81 -> 58.46 | Helped |
| style-based (CBIN) discriminator instead of vanilla conditional (ablation) | DMIT · Yosemite summer->winter | FID: 63.34 -> 58.46 | Helped |
| style-based discriminator instead of projection discriminator (ablation) | DMIT · Yosemite summer->winter | FID: 66.50 -> 58.46 | Helped |
| style-based discriminator instead of projection discriminator (ablation) | DMIT · Yosemite summer->winter | LPIPS: 0.289 -> 0.302 | Mixed |
| unified multi-mapping model vs multi-modal baseline DRIT | DMIT · Yosemite summer->winter | FID: 58.70 -> 58.46 | Helped |
| DMIT vs StarGAN* on facial attribute transfer | DMIT · CelebA | FID: 48.16 -> 32.36 | Helped |
Closest papers by meaning
- Image-to-image translation for cross-domain disentanglement (NeurIPS 2018 · similarity 0.93)
- A Unified Feature Disentangler for Multi-Domain Image Translation and Manipulation (NeurIPS 2018 · similarity 0.90)
- Toward Multimodal Image-to-Image Translation (NeurIPS 2017 · similarity 0.90)
- TraVeLGAN: Image-To-Image Translation by Transformation Vector Learning (CVPR 2019 · similarity 0.90)
- Smoothing the Disentangled Latent Style Space for Unsupervised Image-to-Image Translation (CVPR 2021 · similarity 0.89)
- Image-To-Image Translation With Conditional Adversarial Networks (CVPR 2017 · 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.