Research map / Image generation / Editing and style transfer
WISE: Whitebox Image Stylization by Example-Based Learning
Winfried Lötzsch, Max Reimann, Martin Büssemeyer, Amir Semmo, Jürgen Döllner, Matthias Trapp
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 |
|---|---|---|---|
| train PPN with image-space loss instead of parameter-space loss (l1) (ablation) | global PPN, XDoG · NPR benchmark | SSIM: 0.737 -> 0.780 | Improved |
| train PPN with image-space loss instead of parameter-space loss (l1) (ablation) | global PPN, XDoG · NPR benchmark | PSNR: 12.927 -> 13.875 | Improved |
| image-space loss versus parameter-space loss on ground-truth parameter match (ablation) | global PPN, XDoG · NPR benchmark | parameter l1 error: 0.158 -> 0.183 | Hurt |
| use multi-feature multi-head network instead of ResNet + multi-head (ablation) | global PPN, XDoG · NPR benchmark | SSIM: 0.721 -> 0.780 | Improved |
| add differentiable XDoG with PPN (ResNet post-CNN) over CNN-only (ablation) | APDrawing generator · APDrawing | FID: 62.44 -> 60.55 | Improved |
| add differentiable XDoG with PPN (ResNet post-CNN) over CNN-only (ablation) | APDrawing generator · APDrawing | LPIPS: 0.275 -> 0.285 | No clear effect |
| replace fixed XDoG parameter preset with predicted PPN parameters (ablation) | APDrawing generator · APDrawing | FID: 71.56 -> 60.55 | Improved |
| use ResNet post-processing instead of U-Net (ablation) | APDrawing generator · APDrawing | FID: 89.93 -> 60.55 | Improved |
| add dropout (ablation) | APDrawing generator, PPN + ResNet · APDrawing | FID: 60.55 -> 71.92 | Hurt |
| initialize PPN backbone with ImageNet weights instead of random (ablation) | APDrawing generator, ResNet · APDrawing | FID: 60.55 -> 64.81 | Hurt |
| apply XDoG as post-processing instead of pre-processing (ablation) | APDrawing generator · APDrawing | training convergence | Hurt |
| use PPN + XDoG alone without a post-processing CNN (ablation) | APDrawing generator · APDrawing | training stability | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- APDrawing GAN — APDrawingGAN: Generating Artistic Portrait Drawings From Face Photos With Hierarchical GANs (CVPR 2019) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Rethinking Style Transfer: From Pixels to Parameterized Brushstrokes (CVPR 2021 · similarity 0.89)
- Multimodal Transfer: A Hierarchical Deep Convolutional Neural Network for Fast Artistic Style Transfer (CVPR 2017 · similarity 0.88)
- Adjustable Real-time Style Transfer (ICLR 2020 · similarity 0.88)
- Learning to Stylize by Learning to Destylize: A Scalable Paradigm for Supervised Style Transfer (ECCV 2026 · similarity 0.87)
- StyleMaster: Stylize Your Video with Artistic Generation and Translation (CVPR 2025 · similarity 0.87)
- StyO: Stylize Your Face in Only One-Shot (AAAI 2025 · 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.
