Research map / Image generation / Editing and style transfer
Filter Style Transfer between Photos
Jonghwa Yim, Jisung Yoo, Won-joon Do, Beomsu Kim, Jihwan Choe
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 aleatoric uncertainty weighting to the regression (ablation) | FST · MSCOCO val, 18 Instagram filters | PSNR: 25.226 -> 25.438 | Improved |
| add aleatoric uncertainty weighting to the regression (ablation) | FST · MSCOCO val, 18 Instagram filters | Delta E 2000: 6.660 -> 6.427 | Improved |
| use combined (epistemic + aleatoric) uncertainty weighting (ablation) | FST · MSCOCO val, 18 Instagram filters | PSNR: 25.226 -> 25.495 | Improved |
| use combined uncertainty instead of aleatoric only (ablation) | FST (w. AU) · MSCOCO val, 18 Instagram filters | PSNR: 25.438 -> 25.495 | Improved |
| add regularization terms on top of combined uncertainty (ablation) | FST (w. U) · MSCOCO val, 18 Instagram filters | PSNR: 25.495 -> 26.093 | Improved |
| add regularization terms on top of combined uncertainty (ablation) | FST (w. U) · MSCOCO val, 18 Instagram filters | Delta E 2000: 6.394 -> 6.148 | Improved |
| add channel correlation (CC) term (ablation) | FST (w. U) · MSCOCO val, 18 Instagram filters | PSNR: 25.495 -> 24.931 | Hurt |
| add channel correlation (CC) term (ablation) | FST (w. U) · MSCOCO val, 18 Instagram filters | Delta E 2000: 6.394 -> 6.725 | Hurt |
| use FST instead of WCT2 photorealistic style transfer | WCT2 · MSCOCO val, 18 Instagram filters | PSNR: 16.473 -> 26.093 | Improved |
| use FST instead of linear Color Transfer | Color Transfer · MSCOCO val, 18 Instagram filters | PSNR: 7.325 -> 26.093 | Improved |
| use FST instead of WCT2 on unseen private photos and filters | WCT2 · Private photos, 43 filters | PSNR: 17.234 -> 25.850 | Improved |
| exclude epistemic uncertainty (10 MC dropouts) from FST (ablation) | FST (full) · Run time on PC (filter extraction) | time (ms): 132ms -> 28ms | Improved |
Closest papers by meaning
- CLIPstyler: Image Style Transfer With a Single Text Condition (CVPR 2022 · similarity 0.87)
- Explicit Filterbank Learning for Neural Image Style Transfer and Image Processing (TPAMI 2020 · similarity 0.86)
- A Content Transformation Block for Image Style Transfer (CVPR 2019 · similarity 0.86)
- StyleSSP: Sampling StartPoint Enhancement for Training-free Diffusion-based Method for Style Transfer (CVPR 2025 · similarity 0.86)
- OmniStyle: Filtering High Quality Style Transfer Data at Scale (CVPR 2025 · similarity 0.85)
- Style Injection in Diffusion: A Training-free Approach for Adapting Large-scale Diffusion Models for Style Transfer (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.
