Research map / Image generation / Editing and style transfer
Style-Guided and Disentangled Representation for Robust 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 |
|---|---|---|---|
| Add mode seeking regularization (MS) (ablation) | StarGAN-v2 · Yosemite latent-guided synthesis | FID: 43.5 -> 42.5 | Helped |
| Add mode seeking regularization (MS) (ablation) | StarGAN-v2 · Yosemite latent-guided synthesis | LPIPS: 0.196 -> 0.248 | Helped |
| Add NPMI loss alongside MS (ablation) | StarGAN-v2 + MS · Yosemite latent-guided synthesis | FID: 42.5 -> 41.1 | Helped |
| Add NPMI loss alongside MS (ablation) | StarGAN-v2 + MS · Yosemite latent-guided synthesis | LPIPS: 0.248 -> 0.318 | Helped |
| Add STGD loss (ablation) | StarGAN-v2 + MS + NPMI · Yosemite latent-guided synthesis | FID: 41.1 -> 39.4 | Helped |
| Add STGD loss (ablation) | StarGAN-v2 + MS + NPMI · Yosemite latent-guided synthesis | LPIPS: 0.318 -> 0.372 | Helped |
| Add IG loss (ablation) | StarGAN-v2 + MS + NPMI + STGD · Yosemite latent-guided synthesis | FID: 39.4 -> 39.1 | Helped |
| Add IG loss (ablation) | StarGAN-v2 + MS + NPMI + STGD · Yosemite latent-guided synthesis | LPIPS: 0.372 -> 0.381 | Helped |
| Use SRIT instead of StarGAN-v2 | StarGAN-v2 · AFHQ latent-guided synthesis | FID: 16.2 -> 12.5 | Helped |
| Use SRIT instead of StarGAN-v2 | StarGAN-v2 · AFHQ reference-guided synthesis | LPIPS: 0.432 -> 0.555 | Helped |
Closest papers by meaning
- Smoothing the Disentangled Latent Style Space for Unsupervised Image-to-Image Translation (CVPR 2021 · similarity 0.90)
- Image-to-Image Translation via Hierarchical Style Disentanglement (CVPR 2021 · similarity 0.89)
- StEP: Style-Based Encoder Pre-Training for Multi-Modal Image Synthesis (CVPR 2021 · similarity 0.88)
- A Style-Aware Discriminator for Controllable Image Translation (CVPR 2022 · similarity 0.88)
- Multi-mapping Image-to-Image Translation via Learning Disentanglement (NeurIPS 2019 · similarity 0.87)
- Diffusion-based Image Translation using disentangled style and content representation (ICLR 2023 · 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.