Research map / Image generation / Editing and style transfer
ETNet: Error Transition Network for Arbitrary Style Transfer
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 |
|---|---|---|---|
| increase the number of iterative refinements from one to three (ablation) | ETNet · 100 synthesized images | Style loss: 28.3117 -> 7.1984 | Helped |
| increase the number of iterative refinements from one to three (ablation) | ETNet · 100 synthesized images | Content loss: 10.9117 -> 18.3771 | Hurt |
| replace explicit error features with plain deep features (ablation) | ETNet K=1 · 100 synthesized images | Style loss: 28.3117 -> 28.3513 | Hurt |
| replace explicit error features with plain deep features (ablation) | ETNet K=2 · 100 synthesized images | Style loss: 12.1404 -> 18.3374 | Hurt |
| replace explicit error features with plain deep features (ablation) | ETNet K=3 · 100 synthesized images | Style loss: 7.1984 -> 15.2981 | Hurt |
| add joint analysis of intermediate-stylization features when computing residuals (ablation) | ETNet K'=1 · 100 synthesized images | Style loss: 28.3283 -> 28.3117 | Helped |
| add joint analysis of intermediate-stylization features when computing residuals (ablation) | ETNet K'=2 · 100 synthesized images | Style loss: 23.6618 -> 12.1404 | Helped |
| add joint analysis of intermediate-stylization features when computing residuals (ablation) | ETNet K'=3 · 100 synthesized images | Style loss: 21.3698 -> 7.1984 | Helped |
| use ETNet instead of AdaIN, WCT, Avatar-Net, or AAMS | arbitrary style transfer · User study | Preference (%): 16.1 / 26.4 / 13.0 / 11.2 -> 33.3 | Helped |
| use ETNet instead of AdaIN, WCT, Avatar-Net, or AAMS | arbitrary style transfer · 100 test images | Style loss: 71.5744 / 26.1967 / 42.8833 / 40.2651 -> 7.1984 | Helped |
Closest papers by meaning
- Style-Aware Normalized Loss for Improving Arbitrary Style Transfer (CVPR 2021 · similarity 0.87)
- Rethinking and Improving the Robustness of Image Style Transfer (CVPR 2021 · similarity 0.87)
- OmniStyle: Filtering High Quality Style Transfer Data at Scale (CVPR 2025 · similarity 0.87)
- CSGO: Content-Style Composition in Text-to-Image Generation (NeurIPS 2025 · similarity 0.86)
- OmniConsistency: Learning Style-Agnostic Consistency from Paired Stylization Data (NeurIPS 2025 · similarity 0.85)
- Puff-Net: Efficient Style Transfer with Pure Content and Style Feature Fusion Network (CVPR 2024 · similarity 0.85)
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.