Research map / Image generation / Editing and style transfer
Deep Translation Prior: Test-Time Training for Photorealistic 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 |
|---|---|---|---|
| Use DTP with λc=1/5 instead of SinIR | Photorealistic style transfer · Photorealistic style transfer | PE: 1.00 -> 1.09 | Mixed |
| Use DTP with λc=1/5 instead of SinIR | Photorealistic style transfer · Photorealistic style transfer | SSIM: 0.84 -> 0.82 | No clear effect |
| Use DTP with λc=1/5 instead of Swap. AE | CelebA-HQ · CelebA-HQ | PE: 0.89 -> 0.47 | Helped |
| Use DTP with λc=1/5 instead of SinIR | CelebA-HQ · CelebA-HQ | SSIM: 0.84 -> 0.96 | Helped |
| Use DTP with λc=1/5 instead of Swap. AE | FFHQ · FFHQ | PE: 1.13 -> 1.21 | Hurt |
| Use DTP with λc=1/5 instead of SinIR | FFHQ · FFHQ | SSIM: 0.93 -> 0.98 | Helped |
| Set λc to 1/5 instead of 4/5 (ablation) | DTP · Photorealistic style transfer | PE: 1.36 -> 1.09 | Helped |
| Set λc to 1/5 instead of 4/5 (ablation) | DTP · CelebA-HQ | SSIM: 0.86 -> 0.96 | Helped |
| Set λc to 1/5 instead of 4/5 (ablation) | DTP · FFHQ | PE: 1.4 -> 1.21 | Helped |
| Add warped feature (ablation) | DTP · Ablation user study | Most preferred: 1.8% -> 89.8% | Helped |
| Add warped image (ablation) | DTP · Ablation user study | Most preferred: 3.2% -> 89.8% | Helped |
| Add feature moving average (ablation) | DTP · Ablation user study | Most preferred: 1.5% -> 89.8% | Helped |
Closest papers by meaning
- Deep Photo Style Transfer (CVPR 2017 · similarity 0.90)
- Multimodal Transfer: A Hierarchical Deep Convolutional Neural Network for Fast Artistic Style Transfer (CVPR 2017 · similarity 0.89)
- A Flexible Convolutional Solver for Fast Style Transfers (CVPR 2019 · similarity 0.89)
- Ultrafast Photorealistic Style Transfer via Neural Architecture Search (AAAI 2020 · similarity 0.88)
- Learning Linear Transformations for Fast Image and Video Style Transfer (CVPR 2019 · similarity 0.88)
- Style Injection in Diffusion: A Training-free Approach for Adapting Large-scale Diffusion Models for Style Transfer (CVPR 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.